![Int J Artif Intell ISSN: 2252-8938
A survey of detecting leaf diseases using machine learning and … (Thilagaraj Thangamuthu)
2499
variety of plant diseases. The proposed technique holds the potential for the non-invasive, precise, and
instantaneous diagnosis of plant leaf diseases, facilitating resource allocation and timely interventions. As
precision agriculture gains traction, the combination of ML and DL in disease identification offers a
transformative possibility for resilient and sustainable crop management.
In this paper, we intend to bring out a comprehensive survey of the ML techniques, DL techniques,
and a combination of ML and DL techniques in the early detection of plant leaf diseases. This survey
critically examines the state-of-the-art technologies and expresses informed views and will definitely provide
guidance and ideas for future developments of the application of ML and DL techniques in the early
detection of plant leaf diseases. This will increase the crop yield to achieve food security.
2. LITERATURE REVIEW
The application of ML techniques and DL techniques for early detection of leaf diseases is studied
in the literature. The comparison of the performance of those approaches is presented in Table 1 (see in
Appendix). The table compares methods, algorithms, and accuracy for various leaf diseases.
2.1. Performance of machine learning models
Various ML models are implemented to detect leaf disease [1]–[8]. The support vector machine
(SVM), random forest (RF), linear regression, decision tree (DT), and k-nearest neighbour (KNN) algorithms
exhibited strong performance. The review findings indicate that the SVM method is capable of accurately
detecting and classifying plant diseases with remarkable accuracy. This model undergoes testing on
numerous types of leaves and various diseases affecting leaves.
2.2. Performance of deep learning models
Several methods cited in papers [9]–[35], derived from pre-existing DL models, exhibited robust
performance in this challenge. This study introduces contemporary and superior solutions. The following
deep algorithms shown strong performance in detecting leaf diseases: CNN, Inception V4, DenseNet-121,
ResNet-50, InceptionResNet V2, EResNet-50, VGG16, R-CNN, EfficientNet, GoogleNet, ant colony
optimization with convolution neural network (ACO-CNN), and generative adversarial network (GAN). The
VGG16 algorithm performs well with the highest accuracy. It has been noted that significant deep-learning
algorithms are being used to identify leaf diseases.
2.3. Performance of a combination of machine learning and deep learning models
The integration of ML and DL models was applied in this contest, as described in [36]–[42]. The
performance of the artificial neural network (ANN), KNN, and CNN models was commendable when
handling various leaf datasets containing diverse illnesses. The combined models outperform other
approaches. The combined models perform well with good accuracy in dealing with different leaf datasets
with multiple diseases [43]–[50]. Therefore, it can be said that the combined models outperform ML and DL
models in terms of accuracy when it comes to removing intricate patterns and features from leaf pictures.
3. POTENTIAL AREAS OF RESEARCH
In the previous sections, the authors of this work have identified some opportunities that they believe
have not received much attention from researchers. There is limited discourse on the real-time surveillance of
the initial symptoms of illnesses before their widespread dissemination across the entire plant. There is limited
study on integrating various tracking and testing duties into a single system to lower expenses, enhance
technology accessibility for farmers, and provide ease. Most studies on models for detecting plant diseases
examine two-dimensional images taken from plant samples. When dealing with fruit samples, the use of
single-input cameras or a two-dimensional view can be problematic due to the spherical or cylindrical shape of
most fruits. Hence, there are a lot of research opportunities on sensor technology. Additionally, advances in
DL and ML can be used to create algorithms that more accurately detect leaf illnesses early on.
4. CONCLUSION
The study of DL and ML techniques for the accurate and dependable early detection of plant leaf
diseases is compared in this work. It is discovered that while DL approaches in general and CNN, in
particular, are amazing at learning complex hierarchical representations, using ML techniques for obtaining
and choosing features improves the model's selective abilities. The system is capable of learning hierarchical
properties because DL techniques more particularly, CNN are combined with ML techniques. This allows the
system to be extremely accurate in identifying and classifying various plant leaf diseases. The survey's
findings demonstrate how effective it is to combine DL and ML models for early leaf disease identification.](https://image.slidesharecdn.com/0524961-241223023520-00a5179c/85/A-survey-of-detecting-leaf-diseases-using-machine-learning-and-deep-learning-in-various-crops-2-320.jpg)
![ ISSN: 2252-8938
Int J Artif Intell, Vol. 13, No. 3, September 2024: 2498-2505
2500
APPENDIX
Table 1. Comparison of ML and DL techniques for early detection of leaf diseases (continue…)
Author Work carried
Methods and
algorithms
Plant name and
diseases identified
Accuracy (%)
Ramesh et al.
[1]
The gathered datasets of healthy and diseased leaves are
subjected to a RF collective training procedure. Features
of an image are extracted using the histogram, which is
a representation of an oriented gradient. Large-scale
plant disease detection through the use of ML to train
vast amounts of publicly accessible data sets.
RF, SVM,
KNN, Naïve
Bayes (NB),
CART
Papaya leaf-brown
spot
RF-70.14 %,
SVM-40.33%,
KNN-66.76%,
NB-57.61%,
CART-64.66%
Oo and Htun
[2]
The experimental findings unequivocally establish
that the suggested methodology adeptly discerns and
categorizes four prevalent plant leaf ailments:
cercospora leaf spot, powdery mildew, bacterial
blight, and rust.
SVM, KNN,
and ensemble
classifier
(EC)
Rose-bacterial
blight, powdery
mildew, cercospora
leaf spot and rust
SVM-98.2%,
KNN-80.02%,
EC-84.6%
Mokhtar et al.
[3]
The gabor wavelet change method is used to extract
relevant features from tomato leaf pictures. This is
carried out in conjunction with the application of
different kernel functions through SVM. Finding and
classifying the particular type of illness affecting
tomato plants is the aim.
SVM Tomato leaf-
bacterial blight and
cercospora leaf
spot, powdery
mildew and rust.
SVM-99.5%
Sandhu and
Kaur [4]
These disease detection techniques exhibit high
efficiency and accuracy, enabling them to effectively
operate the created system for leaf disease detection,
despite certain restrictions.
SVM, NB Various plants and
diseases
SVM-86%,
NB-79%
Iqbal and
Talukder [5]
A dataset comprising 450 images of both healthy and
diseased potato leaves is used for image
segmentation. The dataset was sourced from the free-
to-use plant village database. To distinguish between
healthy and diseased leaves, seven classifier
approaches are used.
RF, logistic
regression
(LR), KNN,
DT, NB,
LDA, and
SVM
Potato leaf-late
blight, early blight
RF-97%,
LR-94%,
KNN&DT-91%,
NB-84%,
LDA-78%,
SVM-37%
Islam et al. [6] Multiclass SVM for picture segmentation. SVM Potato leaf-late
blight, early blight
SVM-95%
Ahmed et al.
[7]
Four ML methods have been evaluated and compared
for the detection of illnesses in rice leaves. To
varying degrees, the algorithms correctly anticipated
the rice leaf diseases.
LR, KNN,
DT, NB
Rice leaf-bacterial
blight, brown spot,
leaf smut
DT-97%
Mohan et al.
[8]
The task can be divided into two primary categories:
recognizing paddy disease and signs of plant diseases.
The first step in the disease diagnosis procedure is to
use an AdaBoost classifier and Haar-like features to
identify the rice plant's affected area.
SIFT feature
& classifier-
SVM & KNN
Paddy (rice)-and
multiple diseases
KNN-93.33%,
SVM-91.10%
Andrew et al.
[9]
Plant disease identification efficiency with pre-trained
models based on CNN. Their focus was on optimizing
the hyperparameters of popular pre-trained models such
as ResNet-50, DenseNet-121, VGG-16, and Inception
V4. The well-known PlantVillage dataset, which
includes 54,305 picture samples of various plant disease
species in 38 classifications, was used for the tests.
Inception V4,
DenseNet-
121, ResNet-
50, VGG-16
Apple, tomato, and
grape leaves,
multiple diseases
V4-99.78%,
VGG16-84.27%,
ResNet50-99.83%,
DenseNet121-
99.81%
Sladojevic
et al. [10]
The novel approach to training and the methodology
employed enables a rapid and effortless application of
the system in practical settings. The model that has
been suggested can distinguish between 13 different
types of plant diseases when compared to healthy
leaves. Additionally, it is capable of differentiating
between the surroundings and the leaves of plants.
CNN 13 different plant
leaves
CNN-96.3%
Joshi and
Bhavsar [11]
Detection of nightshade crop leaf disease detection. DL
algorithms
Multiple plants and
diseases
Accuracy not
mentioned
Srinidhi et al.
[12]
The classification is performed with high accuracy,
categorizing the diseases into 4 distinct classes. This
work uses data enrichment and image annotation
techniques, namely flipping, blurring, and canny edge
detection, to improve the apple leaf disease dataset.
CNN-
EfficientNetB
7, DenseNet
Apple leaf-multiple
diseases
EfficientNetB7
-99.8%,
DenseNet-
99.75%
Monigari et al.
[13]
By using 20639 images from 15 folders showing both
healthy and damaged leaves from plants, then CNN
was trained. With the use of leaf photo analysis, this
research aims to develop a more advanced and
accurate method for understanding plant illnesses.
CNN Tomato-late blight,
septoria, pepper-
bacterial spot,
potato-early and
late blight
CNN -90%
Krishnamoorthy
et al. [14]
A CNN model called InceptionResNetV2 is used in
conjunction with the transfer learning technique to
reliably detect diseases in images of rice leaves. By
modifying several hyperparameters, the foundational
CNN model was made more accurate.
CNN model-
Inception
ResNetV2
Rice leaf-leaf blast,
bacterial blight,
and brown spot
Inception
ResNetV2-
95.67%](https://image.slidesharecdn.com/0524961-241223023520-00a5179c/85/A-survey-of-detecting-leaf-diseases-using-machine-learning-and-deep-learning-in-various-crops-3-320.jpg)
![Int J Artif Intell ISSN: 2252-8938
A survey of detecting leaf diseases using machine learning and … (Thilagaraj Thangamuthu)
2501
Table 1. Comparison of ML and DL techniques for early detection of leaf diseases
Author Work carried
Methods and
algorithms
Plant name and
diseases identified
Accuracy (%)
Deng et al.
[15]
The smartphone app, along with a software system
comprising servers and clients, offers precise and
effortless recognition of rice diseases. An inherent
constraint of the ensemble model is the presence of
numerous parameters, which might potentially impact
the speed of identification.
ResNet-50,
ResNeXt-50,
DenseNet-121,
ResNeSt-50, and
SE-ResNet-50
Rice leaf- false
smut, neck blast,
rice leaf blast,
sheath blight,
bacterial stripe,
brown spot
DenseNet-121,
SE-ResNet-50,
and ResNeSt-50
performed well
with accuracy–
98%
Myna et al.
[16]
Healthy and sick leaves are inputs and the afflicted
photos are classified into five distinct categories. The
proposed methodology utilizes various stages,
including preprocessing, feature extraction, training,
testing, and classification.
DL algorithm -
VGG16
Cabbage leaf and
5 diseases
VGG16-92%
Bari et al.
[17]
The faster R-CNN technique incorporates a
sophisticated region proposal network (RPN)
architecture that accurately determines the placement of
objects to produce potential regions. Three different
rice leaf diseases might be automatically diagnosed
with the use of the suggested DL-based method.
Faster R-CNN
algorithm
Rice leaf-rice
blast, brown spot
and hispa
R-CNN for rice
blast–98.09%,
Brown spot–
98.85%, Hispa–
99.17%
Vasantha
et al. [18]
Using a variety of ML techniques and comparing
several algorithms to identify the crop disease kind
based on image data, this system provides possible
solutions. Furthermore, it shows freshly introduced
methodologies and performance measures.
ML and DL
algorithms
Rice leaf-brown
spot, leaf blast,
sheath blight,
sheath rot, bacterial
leaf blight, leaf
Smut
CNN–highest
accuracy
Tejaswini
et al. [19]
This work aids farmers by identifying illnesses in rice
leaves, hence promoting robust crop production.
When it comes to performance, DL models
outperform conventional ML techniques.
5-layer convolution,
VGG-16, VGG19,
Xception, Resnet
RICE leaf- brown
spot, leaf blast,
hispa
5-layer
convolution
model-78.2%
Rajeena et al.
[20]
By modifying the variables, the suggested study uses
EfficientNet to improve the accuracy of the maize
leaf disease database. Tests utilizing DenseNet and
ResNet on the test dataset validate the accuracy and
resilience of this method.
EfficientNet-
based DL
framework
Corn leaf-rust,
gray leaf spot,
blight
EfficientNet-
98.85%
Goyal et al.
[21]
A new DL model that successfully divides wheat
diseases into ten different classifications.
Deep
convolution
architecture
Wheat leaf-8
diseases
Average accuracy
of the proposed
model-98.62%.
Jiang et al.
[22]
The data enrichment and image annotation techniques
are used in this study to construct the apple leaf
disease dataset. The dataset consists of laboratory
photographs as well as complicated images captured
in real-world situations.
SSD, VGG-Net,
INAR-SSD,
CNN
Apple leaf-
alternaria leaf
spot, brown spot,
mosaic, grey spot,
rust
INAR-SSD-
78.80%
Zhang et al.
[23]
Empirical studies demonstrate that recognition
accuracy can be enhanced by augmenting the diversity
of pooling operations, incorporating a rectified linear
unit (ReLU) function and dropout operations, and
iteratively adjusting the model parameters.
GoogLeNet,
Cifar10 model
Maize leaf-
8 diseases
GoogLeNet-
98.9%,
Cifar10-98.8%
Sahasra et al.
[24]
The methodologies utilized to categorize distinct
diseases for leaf disease identification. It also
discusses the pre-processing technique used for the
automatic identification of leaf diseases and the
algorithm used for picture segmentation.
VGG16 Apple and tomato
–bacterial spot,
early and late
blight
VGG16-99.99%
Khan et al.
[25]
The technique of transfer learning is utilized to
initialize the parameters of the suggested deep model.
Data augmentation methods like translation,
rotation, reflection, and scaling were used to reduce
overfitting.
CNN Apple leaf-
8 diseases
CNN-97.18%
Yan et al.
[26]
An improved model that makes use of the VGG16
architecture to precisely identify illnesses in apple leaves
was provided in this study. To reduce the number of
parameters, a pooling layer with global averages is used
in place of the fully connected layer, and a batch
normalization layer is added to accelerate convergence.
CNN based on
VGG-16
Apple leaf-scab,
frog eye spot, and
cedar rust
VGG16-99.01%.
Fuentes et al.
[27]
A technique for annotating classes locally and
globally, as well as augmenting data aiming to
enhance accuracy and minimize the occurrence of
false positives.
Faster R-CNN,
region-based fully
convolutional
network (R-FCN),
and SSD
Tomato leaf–
8 diseases
Not specified
Zhong and
Zhao [28]
Employed ACO-CNN to distinguish between infected
and uninfected leaves.
ACO with
CNN, GAN,
CNN, SGD
Apple leaf-
7 diseases
ACO-CNN–
99.98%,
CNN-99.97%,
GAN-99.6%,
SGD-85%](https://image.slidesharecdn.com/0524961-241223023520-00a5179c/85/A-survey-of-detecting-leaf-diseases-using-machine-learning-and-deep-learning-in-various-crops-4-320.jpg)
![ ISSN: 2252-8938
Int J Artif Intell, Vol. 13, No. 3, September 2024: 2498-2505
2502
Table 1. Comparison of ML and DL techniques for early detection of leaf diseases
Author Work carried
Methods and
algorithms
Plant name and
diseases identified
Accuracy (%)
Malvade
et al. [29]
A novel approach is presented, utilizing pre-trained CNN
models, to automatically detect and categorize biotic
stressors in paddy crops from field photos. An empirical
assessment of the best CNN models employing transfer
learning to learn based on ImageNet weights is also
included in the planned study.
Inception-
V3, VGG-16,
ResNet-50,
DenseNet-
121 and
MobileNet-
28
Paddy (rice)-brown
spot, hip, and leaf
blast
ResNet-50–
92.61%
Mohanty
et al. [30]
DL models can be trained on increasingly bigger, publicly
available image datasets, providing a clear route towards
widespread, smartphone-assisted crop disease diagnosis.
AlexNet,
GoogLeNet
14 crops and 26
diseases
AlexNet-85.53%,
GoogLeNet-
99.34%
Mohameth
et al. [31]
"Smartphone-assisted disease diagnosis" is a breakthrough
made possible by the combination of sophisticated cell
phones with computer vision via DL.
VGG16,
ResNet 50,
Google Net
13 crops and
multiple diseases
VGG16-97.82%,
ResNet50-95.38%,
Google Net-95.3%.
Lee et al.
[32]
A novel method utilizing RNN has been developed to
autonomously identify diseased areas and extract pertinent
characteristics for illness categorization. In addition, they
examine the focal point of attention acquired by our RNN.
InceptionV3,
GoogleNet,
Seq-RNN
Multiple crops and
20 diseases
InceptionV3-
98.05%,
GoogleNet-
99.17%,
Seq-RNN–98.17%
Moganarengam
and Vignesh
[33]
Classification is conducted by analyzing the leaf's
characteristics, such as colour and form, to categorize
diseases into several types, including healthy, bacterial
spot, and leaf mould.
CNN and
DenseNet
201
38 crops and
multiple diseases
DenseNet 201-
95%
Jeon and
Rhee [34]
Using the CNN model, a unique leaf categorization
method was created. Using GoogleNet, two models were
built by changing the network depth. The degree of
damage to the leaves or discoloration was taken into
account to evaluate each model's performance.
GoogleNet,
variant of
GoogleNet
Multiple crops and
diseases
GoogleNet –
99.6 %, variant
of GoogleNet-
99.8%
Sladojevic
et al. [35]
A novel method involving classifying leaf images using
deep neural networks.
CNN 13 crops and
multiple diseases
CNN-96.3%
Sarkar et al.
[36]
A system employing an analysis of colour, edges, and
texture features using SVM and ANN.
SVM and
ANN
Rice leaf - blight SVM-92.4%,
ANN-99.5%
Shrivastava
and Patidar
[37]
This work also looks at the challenges and limitations
associated with using ML and DL to diagnose plant
diseases. These challenges include issues with data
accessibility, imaging quality, and the capacity to
discriminate between plants that are ill and healthy.
ML and DL
algorithms
Multiple datasets
and diseases
ML and DL
algorithms
progress
discussed
Shoaib et al.
[38]
Models for identifying nutritional deficiencies. ANN, SVM,
KNN, and fuzzy
classification
(FC)
Multiple plants and
diseases
ANN-99%,
SVM-97%,
KNN-99 %,
FC-99%
Ngongoma
et al. [39]
Many diseases harm leaves' chlorophyll, which results in dark
or black patches on the leaf's surface. They can be found using
ML techniques, feature extraction, picture preprocessing, and
image segmentation. For feature extraction, the grey level co-
occurrence matrix is employed.
CNN and
SVM
Multiple crops and
diseases
CNN-97.7%,
SVM-80%
Jubaer
et al. [40]
To obtain an accurate diagnosis, linked or related plant
ailments were gathered. The good results obtained with
minimal computing resources demonstrated the algorithm's
efficiency in identifying and categorizing leaf diseases. It is
possible to use more algorithms to improve the
categorization accuracy.
ML and DL
algorithms
Multiple plants and
diseases.
Many
algorithms with
accuracies
Sawarkar and
Kawathekar
[41]
The method of detecting diseases involves acquiring
images, pre-processing images, segmenting pictures,
extracting features, and classifying pictures. Investigating
methods to protect rose plants from various diseases is the
goal of this work.
ML and DL
algorithms
ROSE leaf-black
spot, powdery
mildew,
anthracnose
Recommended
model-SVM
Nikith
et al. [42]
This research examines and presents three distinct models
capable of detecting eight different leaf diseases.
CNN, SVM,
KNN
Soyabean leaf- 7
diseases
CNN-96%,
SVM-76%,
KNN-64%
Singh and
Misra [43]
The subsequent two phases are added in succession after
the segmentation step. The green pixels that predominate
are identified in the first phase. Following this, the green-
dominated pixels are masked using Otsu's approach to
establish the appropriate threshold values.
Proposed
model
Potato leaf– 7
diseases
Proposed
model–94%
Naikwadi
and Amoda
[44]
It includes a summary of the several disease classification
schemes that can be used to find plant leaf diseases. Plant
leaf disease identification relies heavily on image
segmentation, which is accomplished through the use of a
genetic algorithm.
SVM and K-
Means
Rose and beans
leaves-bacterial,
lemon-sun burn,
banana-scorch and
fungal
Proposed
algorithm with
average
accuracy-
97.6%.](https://image.slidesharecdn.com/0524961-241223023520-00a5179c/85/A-survey-of-detecting-leaf-diseases-using-machine-learning-and-deep-learning-in-various-crops-5-320.jpg)
![Int J Artif Intell ISSN: 2252-8938
A survey of detecting leaf diseases using machine learning and … (Thilagaraj Thangamuthu)
2503
Table 1. Comparison of ML and DL techniques for early detection of leaf diseases
Author Work carried
Methods and
algorithms
Plant name and
diseases identified
Accuracy (%)
Kulkarni
et al. [45]
Based on generated data sets, several ML algorithms are
used to discern between wholesome and unwholesome
leaves. The several stages of implementation, including
feature extraction, dataset construction, classifier training,
and classification.
Statistical
image
processing
and ML
model
20 different
diseases of 5
common plants
Proposed
model average
accuracy-93%.
Elfatimi
et al. [46]
Provided a method for classifying leaf diseases in beans and
identifying and describing the optimal network architecture,
including hyperparameters and optimization algorithms.
MobileNet,
MobileNetV2
Beans leaf-
angular leaf spot,
bean rust
Proposed
model 92% to
97%
Bansal
et al. [47]
Presented a collection of pre-trained DL models and
evaluated their efficacy on a dataset comprising
photographs of apple leaves.
CNN Apple leaf-multiple
diseases
Proposed
model-90%
Paymode
and Malode
[48]
Predicting the kind of illness that will affect tomato and grape
leaves in their early stages is the main objective. The multi-
crops leaf disease is detected through the CNN methods.
CNN based
VGG16
model
Tomato and grape-
9 diseases
Proposed model
(grapes-98.40%,
Tomatoes-95.71%)
Orillo et al.
[49]
The effective implementation of a MATLAB programme
involved utilizing image processing and a
backpropagation neural network to accurately identify
illnesses in rice leaves.
ANN Rice leaf -bacterial
leaf blight and rice
blast
Proposed
model-100%
Liu et al.
[50]
To identify diseases in apple leaves, the task entails generating a
sufficient number of abnormal pictures and designing a new
architecture for a deep CNN inspired by AlexNet.
CNN Apple leaf-alternaria
leaf spot, mosaic,
rust, brow spot
Proposed
model-97.62%
REFERENCES
[1] S. Ramesh et al., “Plant disease detection using machine learning,” in 2018 International Conference on Design Innovations for
3Cs Compute Communicate Control (ICDI3C), 2018, pp. 41–45, doi: 10.1109/ICDI3C.2018.00017.
[2] Y. M. Oo and N. C. Htun, “Plant leaf disease detection and classification using image processing,” International Journal of
Research and Engineering, vol. 5, no. 9, pp. 516–523, 2018, doi: 10.21276/ijre.2018.5.9.4.
[3] U. Mokhtar, M. A. S. Ali, A. E. Hassenian, and H. Hefny, “Tomato leaves diseases detection approach based on support vector
machines,” in 2015 11th International Computer Engineering Conference: Today Information Society What’s Next?, 2016, pp.
246–250, doi: 10.1109/ICENCO.2015.7416356.
[4] G. K. Sandhu and R. Kaur, “Plant disease detection techniques: a review,” in 2019 International Conference on Automation,
Computational and Technology Management (ICACTM), 2019, pp. 34–38, doi: 10.1109/ICACTM.2019.8776827.
[5] M. A. Iqbal and K. H. Talukder, “Detection of potato disease using image segmentation and machine learning,” in 2020 International
Conference on Wireless Communications Signal Processing and Networking, 2020, pp. 43–47, doi: 10.1109/WiSPNET48689.2020.9198563.
[6] M. Islam, A. Dinh, K. Wahid, and P. Bhowmik, “Detection of potato diseases using image segmentation and multiclass support vector
machine,”in IEEE 30th Canadian Conference on Electrical and ComputerEngineering, 2017, pp. 1–4, doi: 10.1109/CCECE.2017.7946594.
[7] K. Ahmed, T. R. Shahidi, S. M. I. Alam, and S. Momen, “Rice leaf disease detection using machine learning techniques,” in 2019
International Conference on Sustainable Technologies for Industry 4.0 (STI), 2019, pp. 1–5, doi: 10.1109/STI47673.2019.9068096.
[8] K. J. Mohan, M. Balasubramanian, and S. Palanivel, “Detection and recognition of diseases from paddy plant leaf images,”
International Journal of Computer Applications, vol. 144, no. 12, pp. 34–41, 2016, doi: 10.5120/ijca2016910505.
[9] J. Andrew, J. Eunice, D. E. Popescu, M. K. Chowdary, and J. Hemanth, “Deep learning-based leaf disease detection in crops
using images for agricultural applications,” Agronomy, vol. 12, no. 10, 2022, doi: 10.3390/agronomy12102395.
[10] S. Sladojevic, M. Arsenovic, A. Anderla, D. Culibrk, and D. Stefanovic, “Deep neural networks based recognition of plant diseases by
leaf image classification,” Computational Intelligence and Neuroscience, vol. 2016, pp. 1–12, 2016, doi: 10.1155/2016/3289801.
[11] B. M. Joshi and H. Bhavsar, “Deep learning technology based night-CNN for nightshade crop leaf disease detection,”
International Journal of Intelligent Systems and Applications in Engineering, vol. 11, no. 1, pp. 215–227, 2023.
[12] V. V. Srinidhi, A. Sahay, and K. Deeba, “Plant pathology disease detection in apple leaves using deep convolutional neural
networks apple leaves disease detection using EfficientNet and DenseNet,” in 2021 5th International Conference on Computing
Methodologies and Communication (ICCMC), 2021, pp. 1119–1127, doi: 10.1109/ICCMC51019.2021.9418268.
[13] V. Monigari, G. K. Sri, and T. Prathima, “Plant leaf disease prediction,” International Journal for Research in Applied Science
and Engineering Technology, vol. 9, no. 7, pp. 1295–1305, 2021, doi: 10.22214/ijraset.2021.36582.
[14] N. Krishnamoorthy, L. V. N. Prasad, C. S. P. Kumar, B. Subedi, H. B. Abraha, and S. V. E., “Rice leaf diseases prediction using
deep neural networks with transfer learning,” Environmental Research, vol. 198, 2021, doi: 10.1016/j.envres.2021.111275.
[15] R. Deng et al., “Automatic diagnosis of rice diseases using deep learning,” Frontiers in Plant Science, vol. 12, pp. 1–15, 2021,
doi: 10.3389/fpls.2021.701038.
[16] A. N. Myna, M. K., M. K., R. H. S., and Y. D. Jain, “Classification and detection of cabbage leaf diseases from images using deep
learning methods,” Automation, Control and Intelligent Systems, vol. 11, no. 1, pp. 1–7, 2023, doi: 10.11648/j.acis.20231101.11.
[17] B. S. Bari et al., “A real-time approach of diagnosing rice leaf disease using deep learning-based faster R-CNN framework,”
PeerJ Computer Science, vol. 7, pp. 1–27, 2021, doi: 10.7717/PEERJ-CS.432.
[18] S. V. Vasantha, B. Kiranmai, and S. R. Krishna, “Techniques for rice leaf disease detection using machine learning algorithms,”
International Journal of Engineering and Technical Research, vol. 9, no. 8, pp. 162–166, 2021.
[19] P. Tejaswini, P. Singh, M. Ramchandani, Y. K. Rathore, and R. R. Janghel, “Rice leaf disease classification using CNN,” IOP
Conference Series: Earth and Environmental Science, vol. 1032, no. 1, pp. 1–14, 2022, doi: 10.1088/1755-1315/1032/1/012017.
[20] F. Rajeena P. P, A. S. U, M. A. Moustafa, and M. A. S. Ali, “Detecting plant disease in corn leaf using EfficientNet architecture
an analytical approach,” Electronics, vol. 12, no. 8, pp. 1–18, 2023, doi: 10.3390/electronics12081938.
[21] L. Goyal, C. M. Sharma, A. Singh, and P. K. Singh, “Leaf and spike wheat disease detection & classification using an improved
deep convolutional architecture,” Informatics in Medicine Unlocked, vol. 25, pp. 1–8, 2021, doi: 10.1016/j.imu.2021.100642.
[22] P. Jiang, Y. Chen, B. Liu, D. He, and C. Liang, “Real-time detection of apple leaf diseases using deep learning approach based on
improved convolutional neural networks,” IEEE Access, vol. 7, pp. 59069–59080, 2019, doi: 10.1109/ACCESS.2019.2914929.
[23] X. Zhang, Y. Qiao, F. Meng, C. Fan, and M. Zhang, “Identificbiation of maize leaf diseases using improved deep convolutional](https://image.slidesharecdn.com/0524961-241223023520-00a5179c/85/A-survey-of-detecting-leaf-diseases-using-machine-learning-and-deep-learning-in-various-crops-6-320.jpg)
![ ISSN: 2252-8938
Int J Artif Intell, Vol. 13, No. 3, September 2024: 2498-2505
2504
neural networks,” IEEE Access, vol. 6, pp. 30370–30377, 2018, doi: 10.1109/ACCESS.2018.2844405.
[24] T. Sahasra, S. K. S, S. M. S, and P. R. Devi, “Leaf disease detection using deep learning,” in 2021 Second International Conference on
Electronics and Sustainable Communication Systems (ICESC), 2021, pp. 1797–1804, doi: 10.1109/ICESC51422.2021.9532909.
[25] A. I. Khan, S. M. K. Quadri, S. Banday, and J. L. Shah, “Deep diagnosis: a real-time apple leaf disease detection system based on
deep learning,” Computers and Electronics in Agriculture, vol. 198, 2022, doi: 10.1016/j.compag.2022.107093.
[26] Q. Yan, B. Yang, W. Wang, B. Wang, P. Chen, and J. Zhang, “Apple leaf diseases recognition based on an improved
convolutional neural network,” Sensors, vol. 20, no. 12, pp. 1–14, 2020, doi: 10.3390/s20123535.
[27] A. Fuentes, S. Yoon, S. C. Kim, and D. S. Park, “A robust deep-learning-based detector for real-time tomato plant diseases and
pests recognition,” Sensors, vol. 17, no. 9, pp. 1–21, 2017, doi: 10.3390/s17092022.
[28] Y. Zhong and M. Zhao, “Research on deep learning in apple leaf disease recognition,” Computers and Electronics in Agriculture,
vol. 168, pp. 1–6, 2020, doi: 10.1016/j.compag.2019.105146.
[29] N. N. Malvade, R. Yakkundimath, G. Saunshi, M. C. Elemmi, and P. Baraki, “A comparative analysis of paddy crop biotic stress classification
using pre-trained deep neural networks,” Artificial Intelligence in Agriculture, vol. 6, pp. 167–175, 2022, doi: 10.1016/j.aiia.2022.09.001.
[30] S. P. Mohanty, D. P. Hughes, and M. Salathé, “Using deep learning for image-based plant disease detection,” Frontiers in Plant
Science, vol. 7, pp. 1–10, 2016, doi: 10.3389/fpls.2016.01419.
[31] F. Mohameth, C. Bingcai, and K. A. Sada, “Plant disease detection with deep learning and feature extraction using plant village,”
Journal of Computer and Communications, vol. 08, no. 06, pp. 10–22, 2020, doi: 10.4236/jcc.2020.86002.
[32] S. H. Lee, H. Goëau, P. Bonnet, and A. Joly, “New perspectives on plant disease characterization based on deep learning,”
Computers and Electronics in Agriculture, vol. 170, pp. 1–8, 2020, doi: 10.1016/j.compag.2020.105220.
[33] V. Moganarengam and S. P. Vignesh, “Plant disease classification using CNN and DenseNet, a comparative study,” International
Research Journal of Engineering and Technology (IRJET), vol. 8, no. 11, pp. 236–240, 2021.
[34] W. S. Jeon and S. Y. Rhee, “Plant leaf recognition using a convolution neural network,” International Journal of Fuzzy Logic and
Intelligent Systems, vol. 17, no. 1, pp. 26–34, 2017, doi: 10.5391/IJFIS.2017.17.1.26.
[35] S. Sladojevic, M. Arsenovic, A. Anderla, D. Culibrk, and D. Stefanovic, “Deep neural networks based recognition of plant diseases by
leaf image classification,” Computational Intelligence and Neuroscience, vol. 2016, pp. 1–11, 2016, doi: 10.1155/2016/3289801.
[36] C. Sarkar, D. Gupta, U. Gupta, and B. B. Hazarika, “Leaf disease detection using machine learning and deep learning: review and
challenges,” Applied Soft Computing, vol. 145, 2023, doi: 10.1016/j.asoc.2023.110534.
[37] G. Shrivastava and H. Patidar, “Rice plant disease identification decision support model using machine learning,” ICTACT
Journal on Soft Computing, vol. 12, no. 3, pp. 2619–2627, 2022, doi: 10.21917/ijsc.2022.0365.
[38] M. Shoaib et al., “An advanced deep learning models-based plant disease detection: a review of recent research,” Frontiers in
Plant Science, vol. 14, pp. 1–22, 2023, doi: 10.3389/fpls.2023.1158933.
[39] M. S. P. Ngongoma, M. Kabeya, and K. Moloi, “A review of plant disease detection systems for farming applications,” Applied
Sciences, vol. 13, no. 10, pp. 1–28, 2023, doi: 10.3390/app13105982.
[40] M. A. Jubaer, M. N. Hasan, M. Mustavi, M. T. Shahriar, and T. Ahmed, “Potato leaf disease detection using image processing,”
International Journal of Education and Management Engineering, vol. 13, no. 4, pp. 9–18, 2023, doi: 10.5815/ijeme.2023.04.02.
[41] V. Sawarkar and S. Kawathekar, “A review: rose plant disease detection using image processing,” IOSR Journal of Computer
Engineering (IOSR-JCE), vol. 20, no. 4, pp. 15–19, 2018, doi: 10.9790/0661-2004031519.
[42] B. V. Nikith, N. K. S. Keerthan, M. S. Praneeth, and D. T. Amrita, “Leaf disease detection and classification,” Procedia
Computer Science, vol. 218, pp. 291–300, 2022, doi: 10.1016/j.procs.2023.01.011.
[43] V. Singh and A. K. Misra, “Detection of plant leaf diseases using image segmentation and soft computing techniques,”
Information Processing in Agriculture, vol. 4, no. 1, pp. 41–49, 2017, doi: 10.1016/j.inpa.2016.10.005.
[44] S. Naikwadi and N. Amoda, “Advances in image processing for detection of plant diseases,” International Journal of Application
or Innovation in Engineering & Management (IJAIEM ), vol. 2, no. 11, pp. 168–175, 2013.
[45] P. Kulkarni, A. Karwande, T. Kolhe, S. Kamble, A. Joshi, and M. Wyawahare, “Plant disease detection using image processing
and machine learning,” Computer Vision and Pattern Recognition, pp. 1–13, 2021.
[46] E. Elfatimi, R. Eryigit, and L. Elfatimi, “Beans leaf diseases classification using MobileNet models,” IEEE Access, vol. 10, pp.
9471–9482, 2022, doi: 10.1109/ACCESS.2022.3142817.
[47] P. Bansal, R. Kumar, and S. Kumar, “Disease detection in apple leaves using deep convolutional neural network,” Agriculture,
vol. 11, no. 7, pp. 1–23, 2021, doi: 10.3390/agriculture11070617.
[48] A. S. Paymode and V. B. Malode, “Transfer learning for multi-crop leaf disease image classification using convolutional neural
network VGG,” Artificial Intelligence in Agriculture, vol. 6, pp. 23–33, 2022, doi: 10.1016/j.aiia.2021.12.002.
[49] J. W. Orillo, J. D. Cruz, L. Agapito, P. J. Satimbre, and I. Valenzuela, “Identification of diseases in rice plant (oryza sativa) using back
propagation artificial neural network,” in 2014 International Conference on Humanoid, Nanotechnology, Information Technology,
Communication and Control, Environment and Management (HNICEM), 2014, pp. 1–6, doi: 10.1109/HNICEM.2014.7016248.
[50] B. Liu, Y. Zhang, D. J. He, and Y. Li, “Identification of apple leaf diseases based on deep convolutional neural networks,”
Symmetry, vol. 10, no. 1, pp. 1–16, 2018, doi: 10.3390/sym10010011.
BIOGRAPHIES OF AUTHORS
Dr. Thilagaraj Thangamuthu is currently working as an associate professor and
head in the Department of Artificial Intelligence and Machine Learning, Moodlakatte Institute
of Technology, Kundapura, India. Ph.D. from Bharathiar University, Coimbatore. He has 14
years of Teaching experience; He has published around 12 papers in national and
international conferences and journals. His research areas are machine learning and
educational data mining. He can be contacted at email: thilagaraj.t@gmail.com.](https://image.slidesharecdn.com/0524961-241223023520-00a5179c/85/A-survey-of-detecting-leaf-diseases-using-machine-learning-and-deep-learning-in-various-crops-7-320.jpg)
A survey of detecting leaf diseases using machine learning and deep learning in various crops
- 1. IAES International Journal of Artificial Intelligence (IJ-AI) Vol. 13, No. 3, September 2024, pp. 2498~2505 ISSN: 2252-8938, DOI: 10.11591/ijai.v13.i3.pp2498-2505 2498 Journal homepage: http://ijai.iaescore.com A survey of detecting leaf diseases using machine learning and deep learning in various crops Thilagaraj Thangamuthu1 , Abdul Kareem2 , Varuna Kumara2 , Utkrishna Udesh Naik3 , Sanjana Poojary3 , Bharath Raju3 1 Department of Artificial Intelligence and Machine Learning, Moodlakatte Institute of Technology, Kundapura, India 2 Department of Electronics and Communication Engineering, Moodlakatte Institute of Technology, Kundapura, India 3 Department of Computer Science and Engineering, Moodlakatte Institute of Technology, Kundapura, India Article Info ABSTRACT Article history: Received Jan 9, 2024 Revised Feb 12, 2024 Accepted Mar 4, 2024 For agricultural productivity and food security to be guaranteed, early detection and treatment of illnesses are crucial. Machine learning (ML) and deep learning (DL) approaches can be used to precisely and successfully identify plant leaf diseases. A heterogeneous dataset comprising photos of both healthy and diseased leaves such as bacterial blights, fungal infections, and viral manifestations provides the foundation for the model building and training. Accuracy, precision, recall, and F1-score are the measures used to assess the model's performance. ML techniques are helpful in the identification and extraction of pertinent information from plant leaf pictures, whereas DL techniques in general, and convolutional neural networks (CNN), in particular, are remarkable at learning complex hierarchical representations. Therefore, DL architectures like CNN are utilized in conjunction with ML approaches like support vector machines (SVM), decision trees, and random forests to extract complicated patterns and attributes from leaf pictures. This research provides an extensive analysis of the performance and application of DL and ML approaches recently applied to the early identification of leaf diseases in different crops. Keywords: Agricultural productivity Convolutional neural networks Deep learning Machine learning Plant leaf disease This is an open access article under the CC BY-SA license. Corresponding Author: Thilagaraj Thangamuthu Department of Artificial Intelligence and Machine Learning, Moodlakatte Institute of Technology Kundapura, Karnataka 576217, India Email: thilagaraj@mitkundapura.com 1. INTRODUCTION The global food security is under an increasing threat from widely emerging plant diseases. To increase agricultural productivity and ensure food security, early diagnosis of plant diseases is essential. The conventional methods of detection of plant diseases include visual observation, microscopy, mycological analysis, and biological diagnostics. These methods are of a time-consuming and labor-intensive nature. As a result, there is a lot of research being done on the application of machine learning (ML) and deep learning (DL) approaches to the early identification of plant diseases, and this trend seems highly promising. For the early diagnosis of plant leaf diseases, traditional ML approaches are frequently used for extracting features and classification. DL techniques have recently attracted attention as a potential tool for learning intricate hierarchical representations. Hence, to provide a comprehensive framework for plant leaf disease identification, this study makes use of the synergies between DL and ML techniques. The detection and extraction of relevant information from photographs of plant leaves is made easier by ML techniques. Combining DL and convolutional neural networks (CNN) which excel at learning intricate hierarchical representations creates a potent model that can identify minute patterns that point to a
- 2. Int J Artif Intell ISSN: 2252-8938 A survey of detecting leaf diseases using machine learning and … (Thilagaraj Thangamuthu) 2499 variety of plant diseases. The proposed technique holds the potential for the non-invasive, precise, and instantaneous diagnosis of plant leaf diseases, facilitating resource allocation and timely interventions. As precision agriculture gains traction, the combination of ML and DL in disease identification offers a transformative possibility for resilient and sustainable crop management. In this paper, we intend to bring out a comprehensive survey of the ML techniques, DL techniques, and a combination of ML and DL techniques in the early detection of plant leaf diseases. This survey critically examines the state-of-the-art technologies and expresses informed views and will definitely provide guidance and ideas for future developments of the application of ML and DL techniques in the early detection of plant leaf diseases. This will increase the crop yield to achieve food security. 2. LITERATURE REVIEW The application of ML techniques and DL techniques for early detection of leaf diseases is studied in the literature. The comparison of the performance of those approaches is presented in Table 1 (see in Appendix). The table compares methods, algorithms, and accuracy for various leaf diseases. 2.1. Performance of machine learning models Various ML models are implemented to detect leaf disease [1]–[8]. The support vector machine (SVM), random forest (RF), linear regression, decision tree (DT), and k-nearest neighbour (KNN) algorithms exhibited strong performance. The review findings indicate that the SVM method is capable of accurately detecting and classifying plant diseases with remarkable accuracy. This model undergoes testing on numerous types of leaves and various diseases affecting leaves. 2.2. Performance of deep learning models Several methods cited in papers [9]–[35], derived from pre-existing DL models, exhibited robust performance in this challenge. This study introduces contemporary and superior solutions. The following deep algorithms shown strong performance in detecting leaf diseases: CNN, Inception V4, DenseNet-121, ResNet-50, InceptionResNet V2, EResNet-50, VGG16, R-CNN, EfficientNet, GoogleNet, ant colony optimization with convolution neural network (ACO-CNN), and generative adversarial network (GAN). The VGG16 algorithm performs well with the highest accuracy. It has been noted that significant deep-learning algorithms are being used to identify leaf diseases. 2.3. Performance of a combination of machine learning and deep learning models The integration of ML and DL models was applied in this contest, as described in [36]–[42]. The performance of the artificial neural network (ANN), KNN, and CNN models was commendable when handling various leaf datasets containing diverse illnesses. The combined models outperform other approaches. The combined models perform well with good accuracy in dealing with different leaf datasets with multiple diseases [43]–[50]. Therefore, it can be said that the combined models outperform ML and DL models in terms of accuracy when it comes to removing intricate patterns and features from leaf pictures. 3. POTENTIAL AREAS OF RESEARCH In the previous sections, the authors of this work have identified some opportunities that they believe have not received much attention from researchers. There is limited discourse on the real-time surveillance of the initial symptoms of illnesses before their widespread dissemination across the entire plant. There is limited study on integrating various tracking and testing duties into a single system to lower expenses, enhance technology accessibility for farmers, and provide ease. Most studies on models for detecting plant diseases examine two-dimensional images taken from plant samples. When dealing with fruit samples, the use of single-input cameras or a two-dimensional view can be problematic due to the spherical or cylindrical shape of most fruits. Hence, there are a lot of research opportunities on sensor technology. Additionally, advances in DL and ML can be used to create algorithms that more accurately detect leaf illnesses early on. 4. CONCLUSION The study of DL and ML techniques for the accurate and dependable early detection of plant leaf diseases is compared in this work. It is discovered that while DL approaches in general and CNN, in particular, are amazing at learning complex hierarchical representations, using ML techniques for obtaining and choosing features improves the model's selective abilities. The system is capable of learning hierarchical properties because DL techniques more particularly, CNN are combined with ML techniques. This allows the system to be extremely accurate in identifying and classifying various plant leaf diseases. The survey's findings demonstrate how effective it is to combine DL and ML models for early leaf disease identification.
- 3. ISSN: 2252-8938 Int J Artif Intell, Vol. 13, No. 3, September 2024: 2498-2505 2500 APPENDIX Table 1. Comparison of ML and DL techniques for early detection of leaf diseases (continue…) Author Work carried Methods and algorithms Plant name and diseases identified Accuracy (%) Ramesh et al. [1] The gathered datasets of healthy and diseased leaves are subjected to a RF collective training procedure. Features of an image are extracted using the histogram, which is a representation of an oriented gradient. Large-scale plant disease detection through the use of ML to train vast amounts of publicly accessible data sets. RF, SVM, KNN, Naïve Bayes (NB), CART Papaya leaf-brown spot RF-70.14 %, SVM-40.33%, KNN-66.76%, NB-57.61%, CART-64.66% Oo and Htun [2] The experimental findings unequivocally establish that the suggested methodology adeptly discerns and categorizes four prevalent plant leaf ailments: cercospora leaf spot, powdery mildew, bacterial blight, and rust. SVM, KNN, and ensemble classifier (EC) Rose-bacterial blight, powdery mildew, cercospora leaf spot and rust SVM-98.2%, KNN-80.02%, EC-84.6% Mokhtar et al. [3] The gabor wavelet change method is used to extract relevant features from tomato leaf pictures. This is carried out in conjunction with the application of different kernel functions through SVM. Finding and classifying the particular type of illness affecting tomato plants is the aim. SVM Tomato leaf- bacterial blight and cercospora leaf spot, powdery mildew and rust. SVM-99.5% Sandhu and Kaur [4] These disease detection techniques exhibit high efficiency and accuracy, enabling them to effectively operate the created system for leaf disease detection, despite certain restrictions. SVM, NB Various plants and diseases SVM-86%, NB-79% Iqbal and Talukder [5] A dataset comprising 450 images of both healthy and diseased potato leaves is used for image segmentation. The dataset was sourced from the free- to-use plant village database. To distinguish between healthy and diseased leaves, seven classifier approaches are used. RF, logistic regression (LR), KNN, DT, NB, LDA, and SVM Potato leaf-late blight, early blight RF-97%, LR-94%, KNN&DT-91%, NB-84%, LDA-78%, SVM-37% Islam et al. [6] Multiclass SVM for picture segmentation. SVM Potato leaf-late blight, early blight SVM-95% Ahmed et al. [7] Four ML methods have been evaluated and compared for the detection of illnesses in rice leaves. To varying degrees, the algorithms correctly anticipated the rice leaf diseases. LR, KNN, DT, NB Rice leaf-bacterial blight, brown spot, leaf smut DT-97% Mohan et al. [8] The task can be divided into two primary categories: recognizing paddy disease and signs of plant diseases. The first step in the disease diagnosis procedure is to use an AdaBoost classifier and Haar-like features to identify the rice plant's affected area. SIFT feature & classifier- SVM & KNN Paddy (rice)-and multiple diseases KNN-93.33%, SVM-91.10% Andrew et al. [9] Plant disease identification efficiency with pre-trained models based on CNN. Their focus was on optimizing the hyperparameters of popular pre-trained models such as ResNet-50, DenseNet-121, VGG-16, and Inception V4. The well-known PlantVillage dataset, which includes 54,305 picture samples of various plant disease species in 38 classifications, was used for the tests. Inception V4, DenseNet- 121, ResNet- 50, VGG-16 Apple, tomato, and grape leaves, multiple diseases V4-99.78%, VGG16-84.27%, ResNet50-99.83%, DenseNet121- 99.81% Sladojevic et al. [10] The novel approach to training and the methodology employed enables a rapid and effortless application of the system in practical settings. The model that has been suggested can distinguish between 13 different types of plant diseases when compared to healthy leaves. Additionally, it is capable of differentiating between the surroundings and the leaves of plants. CNN 13 different plant leaves CNN-96.3% Joshi and Bhavsar [11] Detection of nightshade crop leaf disease detection. DL algorithms Multiple plants and diseases Accuracy not mentioned Srinidhi et al. [12] The classification is performed with high accuracy, categorizing the diseases into 4 distinct classes. This work uses data enrichment and image annotation techniques, namely flipping, blurring, and canny edge detection, to improve the apple leaf disease dataset. CNN- EfficientNetB 7, DenseNet Apple leaf-multiple diseases EfficientNetB7 -99.8%, DenseNet- 99.75% Monigari et al. [13] By using 20639 images from 15 folders showing both healthy and damaged leaves from plants, then CNN was trained. With the use of leaf photo analysis, this research aims to develop a more advanced and accurate method for understanding plant illnesses. CNN Tomato-late blight, septoria, pepper- bacterial spot, potato-early and late blight CNN -90% Krishnamoorthy et al. [14] A CNN model called InceptionResNetV2 is used in conjunction with the transfer learning technique to reliably detect diseases in images of rice leaves. By modifying several hyperparameters, the foundational CNN model was made more accurate. CNN model- Inception ResNetV2 Rice leaf-leaf blast, bacterial blight, and brown spot Inception ResNetV2- 95.67%
- 4. Int J Artif Intell ISSN: 2252-8938 A survey of detecting leaf diseases using machine learning and … (Thilagaraj Thangamuthu) 2501 Table 1. Comparison of ML and DL techniques for early detection of leaf diseases Author Work carried Methods and algorithms Plant name and diseases identified Accuracy (%) Deng et al. [15] The smartphone app, along with a software system comprising servers and clients, offers precise and effortless recognition of rice diseases. An inherent constraint of the ensemble model is the presence of numerous parameters, which might potentially impact the speed of identification. ResNet-50, ResNeXt-50, DenseNet-121, ResNeSt-50, and SE-ResNet-50 Rice leaf- false smut, neck blast, rice leaf blast, sheath blight, bacterial stripe, brown spot DenseNet-121, SE-ResNet-50, and ResNeSt-50 performed well with accuracy– 98% Myna et al. [16] Healthy and sick leaves are inputs and the afflicted photos are classified into five distinct categories. The proposed methodology utilizes various stages, including preprocessing, feature extraction, training, testing, and classification. DL algorithm - VGG16 Cabbage leaf and 5 diseases VGG16-92% Bari et al. [17] The faster R-CNN technique incorporates a sophisticated region proposal network (RPN) architecture that accurately determines the placement of objects to produce potential regions. Three different rice leaf diseases might be automatically diagnosed with the use of the suggested DL-based method. Faster R-CNN algorithm Rice leaf-rice blast, brown spot and hispa R-CNN for rice blast–98.09%, Brown spot– 98.85%, Hispa– 99.17% Vasantha et al. [18] Using a variety of ML techniques and comparing several algorithms to identify the crop disease kind based on image data, this system provides possible solutions. Furthermore, it shows freshly introduced methodologies and performance measures. ML and DL algorithms Rice leaf-brown spot, leaf blast, sheath blight, sheath rot, bacterial leaf blight, leaf Smut CNN–highest accuracy Tejaswini et al. [19] This work aids farmers by identifying illnesses in rice leaves, hence promoting robust crop production. When it comes to performance, DL models outperform conventional ML techniques. 5-layer convolution, VGG-16, VGG19, Xception, Resnet RICE leaf- brown spot, leaf blast, hispa 5-layer convolution model-78.2% Rajeena et al. [20] By modifying the variables, the suggested study uses EfficientNet to improve the accuracy of the maize leaf disease database. Tests utilizing DenseNet and ResNet on the test dataset validate the accuracy and resilience of this method. EfficientNet- based DL framework Corn leaf-rust, gray leaf spot, blight EfficientNet- 98.85% Goyal et al. [21] A new DL model that successfully divides wheat diseases into ten different classifications. Deep convolution architecture Wheat leaf-8 diseases Average accuracy of the proposed model-98.62%. Jiang et al. [22] The data enrichment and image annotation techniques are used in this study to construct the apple leaf disease dataset. The dataset consists of laboratory photographs as well as complicated images captured in real-world situations. SSD, VGG-Net, INAR-SSD, CNN Apple leaf- alternaria leaf spot, brown spot, mosaic, grey spot, rust INAR-SSD- 78.80% Zhang et al. [23] Empirical studies demonstrate that recognition accuracy can be enhanced by augmenting the diversity of pooling operations, incorporating a rectified linear unit (ReLU) function and dropout operations, and iteratively adjusting the model parameters. GoogLeNet, Cifar10 model Maize leaf- 8 diseases GoogLeNet- 98.9%, Cifar10-98.8% Sahasra et al. [24] The methodologies utilized to categorize distinct diseases for leaf disease identification. It also discusses the pre-processing technique used for the automatic identification of leaf diseases and the algorithm used for picture segmentation. VGG16 Apple and tomato –bacterial spot, early and late blight VGG16-99.99% Khan et al. [25] The technique of transfer learning is utilized to initialize the parameters of the suggested deep model. Data augmentation methods like translation, rotation, reflection, and scaling were used to reduce overfitting. CNN Apple leaf- 8 diseases CNN-97.18% Yan et al. [26] An improved model that makes use of the VGG16 architecture to precisely identify illnesses in apple leaves was provided in this study. To reduce the number of parameters, a pooling layer with global averages is used in place of the fully connected layer, and a batch normalization layer is added to accelerate convergence. CNN based on VGG-16 Apple leaf-scab, frog eye spot, and cedar rust VGG16-99.01%. Fuentes et al. [27] A technique for annotating classes locally and globally, as well as augmenting data aiming to enhance accuracy and minimize the occurrence of false positives. Faster R-CNN, region-based fully convolutional network (R-FCN), and SSD Tomato leaf– 8 diseases Not specified Zhong and Zhao [28] Employed ACO-CNN to distinguish between infected and uninfected leaves. ACO with CNN, GAN, CNN, SGD Apple leaf- 7 diseases ACO-CNN– 99.98%, CNN-99.97%, GAN-99.6%, SGD-85%
- 5. ISSN: 2252-8938 Int J Artif Intell, Vol. 13, No. 3, September 2024: 2498-2505 2502 Table 1. Comparison of ML and DL techniques for early detection of leaf diseases Author Work carried Methods and algorithms Plant name and diseases identified Accuracy (%) Malvade et al. [29] A novel approach is presented, utilizing pre-trained CNN models, to automatically detect and categorize biotic stressors in paddy crops from field photos. An empirical assessment of the best CNN models employing transfer learning to learn based on ImageNet weights is also included in the planned study. Inception- V3, VGG-16, ResNet-50, DenseNet- 121 and MobileNet- 28 Paddy (rice)-brown spot, hip, and leaf blast ResNet-50– 92.61% Mohanty et al. [30] DL models can be trained on increasingly bigger, publicly available image datasets, providing a clear route towards widespread, smartphone-assisted crop disease diagnosis. AlexNet, GoogLeNet 14 crops and 26 diseases AlexNet-85.53%, GoogLeNet- 99.34% Mohameth et al. [31] "Smartphone-assisted disease diagnosis" is a breakthrough made possible by the combination of sophisticated cell phones with computer vision via DL. VGG16, ResNet 50, Google Net 13 crops and multiple diseases VGG16-97.82%, ResNet50-95.38%, Google Net-95.3%. Lee et al. [32] A novel method utilizing RNN has been developed to autonomously identify diseased areas and extract pertinent characteristics for illness categorization. In addition, they examine the focal point of attention acquired by our RNN. InceptionV3, GoogleNet, Seq-RNN Multiple crops and 20 diseases InceptionV3- 98.05%, GoogleNet- 99.17%, Seq-RNN–98.17% Moganarengam and Vignesh [33] Classification is conducted by analyzing the leaf's characteristics, such as colour and form, to categorize diseases into several types, including healthy, bacterial spot, and leaf mould. CNN and DenseNet 201 38 crops and multiple diseases DenseNet 201- 95% Jeon and Rhee [34] Using the CNN model, a unique leaf categorization method was created. Using GoogleNet, two models were built by changing the network depth. The degree of damage to the leaves or discoloration was taken into account to evaluate each model's performance. GoogleNet, variant of GoogleNet Multiple crops and diseases GoogleNet – 99.6 %, variant of GoogleNet- 99.8% Sladojevic et al. [35] A novel method involving classifying leaf images using deep neural networks. CNN 13 crops and multiple diseases CNN-96.3% Sarkar et al. [36] A system employing an analysis of colour, edges, and texture features using SVM and ANN. SVM and ANN Rice leaf - blight SVM-92.4%, ANN-99.5% Shrivastava and Patidar [37] This work also looks at the challenges and limitations associated with using ML and DL to diagnose plant diseases. These challenges include issues with data accessibility, imaging quality, and the capacity to discriminate between plants that are ill and healthy. ML and DL algorithms Multiple datasets and diseases ML and DL algorithms progress discussed Shoaib et al. [38] Models for identifying nutritional deficiencies. ANN, SVM, KNN, and fuzzy classification (FC) Multiple plants and diseases ANN-99%, SVM-97%, KNN-99 %, FC-99% Ngongoma et al. [39] Many diseases harm leaves' chlorophyll, which results in dark or black patches on the leaf's surface. They can be found using ML techniques, feature extraction, picture preprocessing, and image segmentation. For feature extraction, the grey level co- occurrence matrix is employed. CNN and SVM Multiple crops and diseases CNN-97.7%, SVM-80% Jubaer et al. [40] To obtain an accurate diagnosis, linked or related plant ailments were gathered. The good results obtained with minimal computing resources demonstrated the algorithm's efficiency in identifying and categorizing leaf diseases. It is possible to use more algorithms to improve the categorization accuracy. ML and DL algorithms Multiple plants and diseases. Many algorithms with accuracies Sawarkar and Kawathekar [41] The method of detecting diseases involves acquiring images, pre-processing images, segmenting pictures, extracting features, and classifying pictures. Investigating methods to protect rose plants from various diseases is the goal of this work. ML and DL algorithms ROSE leaf-black spot, powdery mildew, anthracnose Recommended model-SVM Nikith et al. [42] This research examines and presents three distinct models capable of detecting eight different leaf diseases. CNN, SVM, KNN Soyabean leaf- 7 diseases CNN-96%, SVM-76%, KNN-64% Singh and Misra [43] The subsequent two phases are added in succession after the segmentation step. The green pixels that predominate are identified in the first phase. Following this, the green- dominated pixels are masked using Otsu's approach to establish the appropriate threshold values. Proposed model Potato leaf– 7 diseases Proposed model–94% Naikwadi and Amoda [44] It includes a summary of the several disease classification schemes that can be used to find plant leaf diseases. Plant leaf disease identification relies heavily on image segmentation, which is accomplished through the use of a genetic algorithm. SVM and K- Means Rose and beans leaves-bacterial, lemon-sun burn, banana-scorch and fungal Proposed algorithm with average accuracy- 97.6%.
- 6. Int J Artif Intell ISSN: 2252-8938 A survey of detecting leaf diseases using machine learning and … (Thilagaraj Thangamuthu) 2503 Table 1. Comparison of ML and DL techniques for early detection of leaf diseases Author Work carried Methods and algorithms Plant name and diseases identified Accuracy (%) Kulkarni et al. [45] Based on generated data sets, several ML algorithms are used to discern between wholesome and unwholesome leaves. The several stages of implementation, including feature extraction, dataset construction, classifier training, and classification. Statistical image processing and ML model 20 different diseases of 5 common plants Proposed model average accuracy-93%. Elfatimi et al. [46] Provided a method for classifying leaf diseases in beans and identifying and describing the optimal network architecture, including hyperparameters and optimization algorithms. MobileNet, MobileNetV2 Beans leaf- angular leaf spot, bean rust Proposed model 92% to 97% Bansal et al. [47] Presented a collection of pre-trained DL models and evaluated their efficacy on a dataset comprising photographs of apple leaves. CNN Apple leaf-multiple diseases Proposed model-90% Paymode and Malode [48] Predicting the kind of illness that will affect tomato and grape leaves in their early stages is the main objective. The multi- crops leaf disease is detected through the CNN methods. CNN based VGG16 model Tomato and grape- 9 diseases Proposed model (grapes-98.40%, Tomatoes-95.71%) Orillo et al. [49] The effective implementation of a MATLAB programme involved utilizing image processing and a backpropagation neural network to accurately identify illnesses in rice leaves. ANN Rice leaf -bacterial leaf blight and rice blast Proposed model-100% Liu et al. [50] To identify diseases in apple leaves, the task entails generating a sufficient number of abnormal pictures and designing a new architecture for a deep CNN inspired by AlexNet. CNN Apple leaf-alternaria leaf spot, mosaic, rust, brow spot Proposed model-97.62% REFERENCES [1] S. Ramesh et al., “Plant disease detection using machine learning,” in 2018 International Conference on Design Innovations for 3Cs Compute Communicate Control (ICDI3C), 2018, pp. 41–45, doi: 10.1109/ICDI3C.2018.00017. [2] Y. M. Oo and N. C. Htun, “Plant leaf disease detection and classification using image processing,” International Journal of Research and Engineering, vol. 5, no. 9, pp. 516–523, 2018, doi: 10.21276/ijre.2018.5.9.4. [3] U. Mokhtar, M. A. S. Ali, A. E. Hassenian, and H. Hefny, “Tomato leaves diseases detection approach based on support vector machines,” in 2015 11th International Computer Engineering Conference: Today Information Society What’s Next?, 2016, pp. 246–250, doi: 10.1109/ICENCO.2015.7416356. [4] G. K. Sandhu and R. Kaur, “Plant disease detection techniques: a review,” in 2019 International Conference on Automation, Computational and Technology Management (ICACTM), 2019, pp. 34–38, doi: 10.1109/ICACTM.2019.8776827. [5] M. A. Iqbal and K. H. Talukder, “Detection of potato disease using image segmentation and machine learning,” in 2020 International Conference on Wireless Communications Signal Processing and Networking, 2020, pp. 43–47, doi: 10.1109/WiSPNET48689.2020.9198563. [6] M. Islam, A. Dinh, K. Wahid, and P. Bhowmik, “Detection of potato diseases using image segmentation and multiclass support vector machine,”in IEEE 30th Canadian Conference on Electrical and ComputerEngineering, 2017, pp. 1–4, doi: 10.1109/CCECE.2017.7946594. [7] K. Ahmed, T. R. Shahidi, S. M. I. Alam, and S. Momen, “Rice leaf disease detection using machine learning techniques,” in 2019 International Conference on Sustainable Technologies for Industry 4.0 (STI), 2019, pp. 1–5, doi: 10.1109/STI47673.2019.9068096. [8] K. J. Mohan, M. Balasubramanian, and S. Palanivel, “Detection and recognition of diseases from paddy plant leaf images,” International Journal of Computer Applications, vol. 144, no. 12, pp. 34–41, 2016, doi: 10.5120/ijca2016910505. [9] J. Andrew, J. Eunice, D. E. Popescu, M. K. Chowdary, and J. Hemanth, “Deep learning-based leaf disease detection in crops using images for agricultural applications,” Agronomy, vol. 12, no. 10, 2022, doi: 10.3390/agronomy12102395. [10] S. Sladojevic, M. Arsenovic, A. Anderla, D. Culibrk, and D. Stefanovic, “Deep neural networks based recognition of plant diseases by leaf image classification,” Computational Intelligence and Neuroscience, vol. 2016, pp. 1–12, 2016, doi: 10.1155/2016/3289801. [11] B. M. Joshi and H. Bhavsar, “Deep learning technology based night-CNN for nightshade crop leaf disease detection,” International Journal of Intelligent Systems and Applications in Engineering, vol. 11, no. 1, pp. 215–227, 2023. [12] V. V. Srinidhi, A. Sahay, and K. Deeba, “Plant pathology disease detection in apple leaves using deep convolutional neural networks apple leaves disease detection using EfficientNet and DenseNet,” in 2021 5th International Conference on Computing Methodologies and Communication (ICCMC), 2021, pp. 1119–1127, doi: 10.1109/ICCMC51019.2021.9418268. [13] V. Monigari, G. K. Sri, and T. Prathima, “Plant leaf disease prediction,” International Journal for Research in Applied Science and Engineering Technology, vol. 9, no. 7, pp. 1295–1305, 2021, doi: 10.22214/ijraset.2021.36582. [14] N. Krishnamoorthy, L. V. N. Prasad, C. S. P. Kumar, B. Subedi, H. B. Abraha, and S. V. E., “Rice leaf diseases prediction using deep neural networks with transfer learning,” Environmental Research, vol. 198, 2021, doi: 10.1016/j.envres.2021.111275. [15] R. Deng et al., “Automatic diagnosis of rice diseases using deep learning,” Frontiers in Plant Science, vol. 12, pp. 1–15, 2021, doi: 10.3389/fpls.2021.701038. [16] A. N. Myna, M. K., M. K., R. H. S., and Y. D. Jain, “Classification and detection of cabbage leaf diseases from images using deep learning methods,” Automation, Control and Intelligent Systems, vol. 11, no. 1, pp. 1–7, 2023, doi: 10.11648/j.acis.20231101.11. [17] B. S. Bari et al., “A real-time approach of diagnosing rice leaf disease using deep learning-based faster R-CNN framework,” PeerJ Computer Science, vol. 7, pp. 1–27, 2021, doi: 10.7717/PEERJ-CS.432. [18] S. V. Vasantha, B. Kiranmai, and S. R. Krishna, “Techniques for rice leaf disease detection using machine learning algorithms,” International Journal of Engineering and Technical Research, vol. 9, no. 8, pp. 162–166, 2021. [19] P. Tejaswini, P. Singh, M. Ramchandani, Y. K. Rathore, and R. R. Janghel, “Rice leaf disease classification using CNN,” IOP Conference Series: Earth and Environmental Science, vol. 1032, no. 1, pp. 1–14, 2022, doi: 10.1088/1755-1315/1032/1/012017. [20] F. Rajeena P. P, A. S. U, M. A. Moustafa, and M. A. S. Ali, “Detecting plant disease in corn leaf using EfficientNet architecture an analytical approach,” Electronics, vol. 12, no. 8, pp. 1–18, 2023, doi: 10.3390/electronics12081938. [21] L. Goyal, C. M. Sharma, A. Singh, and P. K. Singh, “Leaf and spike wheat disease detection & classification using an improved deep convolutional architecture,” Informatics in Medicine Unlocked, vol. 25, pp. 1–8, 2021, doi: 10.1016/j.imu.2021.100642. [22] P. Jiang, Y. Chen, B. Liu, D. He, and C. Liang, “Real-time detection of apple leaf diseases using deep learning approach based on improved convolutional neural networks,” IEEE Access, vol. 7, pp. 59069–59080, 2019, doi: 10.1109/ACCESS.2019.2914929. [23] X. Zhang, Y. Qiao, F. Meng, C. Fan, and M. Zhang, “Identificbiation of maize leaf diseases using improved deep convolutional
- 7. ISSN: 2252-8938 Int J Artif Intell, Vol. 13, No. 3, September 2024: 2498-2505 2504 neural networks,” IEEE Access, vol. 6, pp. 30370–30377, 2018, doi: 10.1109/ACCESS.2018.2844405. [24] T. Sahasra, S. K. S, S. M. S, and P. R. Devi, “Leaf disease detection using deep learning,” in 2021 Second International Conference on Electronics and Sustainable Communication Systems (ICESC), 2021, pp. 1797–1804, doi: 10.1109/ICESC51422.2021.9532909. [25] A. I. Khan, S. M. K. Quadri, S. Banday, and J. L. Shah, “Deep diagnosis: a real-time apple leaf disease detection system based on deep learning,” Computers and Electronics in Agriculture, vol. 198, 2022, doi: 10.1016/j.compag.2022.107093. [26] Q. Yan, B. Yang, W. Wang, B. Wang, P. Chen, and J. Zhang, “Apple leaf diseases recognition based on an improved convolutional neural network,” Sensors, vol. 20, no. 12, pp. 1–14, 2020, doi: 10.3390/s20123535. [27] A. Fuentes, S. Yoon, S. C. Kim, and D. S. Park, “A robust deep-learning-based detector for real-time tomato plant diseases and pests recognition,” Sensors, vol. 17, no. 9, pp. 1–21, 2017, doi: 10.3390/s17092022. [28] Y. Zhong and M. Zhao, “Research on deep learning in apple leaf disease recognition,” Computers and Electronics in Agriculture, vol. 168, pp. 1–6, 2020, doi: 10.1016/j.compag.2019.105146. [29] N. N. Malvade, R. Yakkundimath, G. Saunshi, M. C. Elemmi, and P. Baraki, “A comparative analysis of paddy crop biotic stress classification using pre-trained deep neural networks,” Artificial Intelligence in Agriculture, vol. 6, pp. 167–175, 2022, doi: 10.1016/j.aiia.2022.09.001. [30] S. P. Mohanty, D. P. Hughes, and M. Salathé, “Using deep learning for image-based plant disease detection,” Frontiers in Plant Science, vol. 7, pp. 1–10, 2016, doi: 10.3389/fpls.2016.01419. [31] F. Mohameth, C. Bingcai, and K. A. Sada, “Plant disease detection with deep learning and feature extraction using plant village,” Journal of Computer and Communications, vol. 08, no. 06, pp. 10–22, 2020, doi: 10.4236/jcc.2020.86002. [32] S. H. Lee, H. Goëau, P. Bonnet, and A. Joly, “New perspectives on plant disease characterization based on deep learning,” Computers and Electronics in Agriculture, vol. 170, pp. 1–8, 2020, doi: 10.1016/j.compag.2020.105220. [33] V. Moganarengam and S. P. Vignesh, “Plant disease classification using CNN and DenseNet, a comparative study,” International Research Journal of Engineering and Technology (IRJET), vol. 8, no. 11, pp. 236–240, 2021. [34] W. S. Jeon and S. Y. Rhee, “Plant leaf recognition using a convolution neural network,” International Journal of Fuzzy Logic and Intelligent Systems, vol. 17, no. 1, pp. 26–34, 2017, doi: 10.5391/IJFIS.2017.17.1.26. [35] S. Sladojevic, M. Arsenovic, A. Anderla, D. Culibrk, and D. Stefanovic, “Deep neural networks based recognition of plant diseases by leaf image classification,” Computational Intelligence and Neuroscience, vol. 2016, pp. 1–11, 2016, doi: 10.1155/2016/3289801. [36] C. Sarkar, D. Gupta, U. Gupta, and B. B. Hazarika, “Leaf disease detection using machine learning and deep learning: review and challenges,” Applied Soft Computing, vol. 145, 2023, doi: 10.1016/j.asoc.2023.110534. [37] G. Shrivastava and H. Patidar, “Rice plant disease identification decision support model using machine learning,” ICTACT Journal on Soft Computing, vol. 12, no. 3, pp. 2619–2627, 2022, doi: 10.21917/ijsc.2022.0365. [38] M. Shoaib et al., “An advanced deep learning models-based plant disease detection: a review of recent research,” Frontiers in Plant Science, vol. 14, pp. 1–22, 2023, doi: 10.3389/fpls.2023.1158933. [39] M. S. P. Ngongoma, M. Kabeya, and K. Moloi, “A review of plant disease detection systems for farming applications,” Applied Sciences, vol. 13, no. 10, pp. 1–28, 2023, doi: 10.3390/app13105982. [40] M. A. Jubaer, M. N. Hasan, M. Mustavi, M. T. Shahriar, and T. Ahmed, “Potato leaf disease detection using image processing,” International Journal of Education and Management Engineering, vol. 13, no. 4, pp. 9–18, 2023, doi: 10.5815/ijeme.2023.04.02. [41] V. Sawarkar and S. Kawathekar, “A review: rose plant disease detection using image processing,” IOSR Journal of Computer Engineering (IOSR-JCE), vol. 20, no. 4, pp. 15–19, 2018, doi: 10.9790/0661-2004031519. [42] B. V. Nikith, N. K. S. Keerthan, M. S. Praneeth, and D. T. Amrita, “Leaf disease detection and classification,” Procedia Computer Science, vol. 218, pp. 291–300, 2022, doi: 10.1016/j.procs.2023.01.011. [43] V. Singh and A. K. Misra, “Detection of plant leaf diseases using image segmentation and soft computing techniques,” Information Processing in Agriculture, vol. 4, no. 1, pp. 41–49, 2017, doi: 10.1016/j.inpa.2016.10.005. [44] S. Naikwadi and N. Amoda, “Advances in image processing for detection of plant diseases,” International Journal of Application or Innovation in Engineering & Management (IJAIEM ), vol. 2, no. 11, pp. 168–175, 2013. [45] P. Kulkarni, A. Karwande, T. Kolhe, S. Kamble, A. Joshi, and M. Wyawahare, “Plant disease detection using image processing and machine learning,” Computer Vision and Pattern Recognition, pp. 1–13, 2021. [46] E. Elfatimi, R. Eryigit, and L. Elfatimi, “Beans leaf diseases classification using MobileNet models,” IEEE Access, vol. 10, pp. 9471–9482, 2022, doi: 10.1109/ACCESS.2022.3142817. [47] P. Bansal, R. Kumar, and S. Kumar, “Disease detection in apple leaves using deep convolutional neural network,” Agriculture, vol. 11, no. 7, pp. 1–23, 2021, doi: 10.3390/agriculture11070617. [48] A. S. Paymode and V. B. Malode, “Transfer learning for multi-crop leaf disease image classification using convolutional neural network VGG,” Artificial Intelligence in Agriculture, vol. 6, pp. 23–33, 2022, doi: 10.1016/j.aiia.2021.12.002. [49] J. W. Orillo, J. D. Cruz, L. Agapito, P. J. Satimbre, and I. Valenzuela, “Identification of diseases in rice plant (oryza sativa) using back propagation artificial neural network,” in 2014 International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment and Management (HNICEM), 2014, pp. 1–6, doi: 10.1109/HNICEM.2014.7016248. [50] B. Liu, Y. Zhang, D. J. He, and Y. Li, “Identification of apple leaf diseases based on deep convolutional neural networks,” Symmetry, vol. 10, no. 1, pp. 1–16, 2018, doi: 10.3390/sym10010011. BIOGRAPHIES OF AUTHORS Dr. Thilagaraj Thangamuthu is currently working as an associate professor and head in the Department of Artificial Intelligence and Machine Learning, Moodlakatte Institute of Technology, Kundapura, India. Ph.D. from Bharathiar University, Coimbatore. He has 14 years of Teaching experience; He has published around 12 papers in national and international conferences and journals. His research areas are machine learning and educational data mining. He can be contacted at email: thilagaraj.t@gmail.com.
- 8. Int J Artif Intell ISSN: 2252-8938 A survey of detecting leaf diseases using machine learning and … (Thilagaraj Thangamuthu) 2505 Dr. Abdul Kareem holds a Doctor of Philosophy from St Peter’s Institute of Higher Education and Research, Chennai, India. He also received his B.Tech. and M.Tech. from Kannur University, India in 2003 and Visvesvaraya Technological University, Belagavi, India in 2008 respectively. He is currently the Principal and a Professor of Electronics and Communication Engineering at Moodlakatte Institute of Technology, Kundapura, India. His research interests are in artificial intelligence, machine learning, control systems, and microelectronics. He has published over 15 papers in international journals and conferences. He is a senior member of IEEE. He can be contacted at email: afthabakareem@gmail.com. Varuna Kumara is a research scholar in the Department of Electronics Engineering at JAIN Deemed to be University, Bengaluru, India. He also received his B.E. and M.Tech. from VTU, Belagavi, India. He is currently serving as an assistant professor of Electronics and Communication Engineering at Moodlakatte Institute of Technology, Kundapura, India. His research interests are in artificial intelligence, signal processing, and control systems. He can be contacted at email: varunakumara@mitkundapura.com. Utkrishna Udesh Naik is currently a student, pursuing his final year of Engineering in the Department of Computer Science and Engineering at Moodlakatte Institute of Technology, Moodlakatte, Kundapura, Udupi, Karnataka, India. He is an active researcher in machine learning and applications. He is a student member of IEEE. His research interests include artificial intelligence, machine learning, and data mining. He can be contacted at email: tkrishnanaik416@gmail.com, utkrishnanaik416@gmail.com. Sanjana Poojary is currently a student, pursuing her final year of Engineering in the Department of Computer Science and Engineering at Moodlakatte Institute of Technology, Moodlakatte, Kundapura, Udupi, Karnataka, India. She is an active researcher in deep learning and applications. She is the student research coordinator of Moodlakatte Institute of Technology. Her research interests include artificial intelligence, machine learning, and fuzzy logic. She can be contacted at email: sanjanapoojaryg@gmail.com. Bharath Raju is currently a student, pursuing his final year of engineering in the Department of Computer Science and Engineering at Moodlakatte Institute of Technology, Moodlakatte, Kundapura, Udupi, Karnataka, India. He is an active researcher in deep learning and applications. His research interests include artificial intelligence, machine learning, and support vector machines. He can be contacted at email: bharath.r32@outlook.com.