Performance Comparison Analysis for Medical Images Using Deep Learning Approaches

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 07 | July 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1046
Performance Comparison Analysis for Medical Images Using Deep
Learning Approaches
Neha Adawadkar, Roshani Chavan, Chaitra Patwardhan, Advait Thakur, Prof. Dr. Mrs. Suhasini
Itkar
Neha Adawadkar, Dept. of Computer Engineering, PES Modern College of Engineering, Maharashtra, India
Roshani Chavan, Dept. of Computer Engineering, PES Modern College of Engineering, Maharashtra, India
Chaitra Patwardhan, Dept. of Computer Engineering, PES Modern College of Engineering, Maharashtra, India
Advait Thakur, Dept. of Computer Engineering, PES Modern College of Engineering, Maharashtra, India
Prof. Dr. Mrs. Suhasini Itkar, Dept. of Computer Engineering, PES Modern College of Engineering, Maharashtra,
India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The recent pandemic has enforced a lot of stress
on our healthcare industry as well as the increase in
population has brought to light that the work bestowed upon
the healthcare specialists needs to be reduced. Deep learning
models such as neural networks are efficient in finding hidden
patterns, which assists the experts in the specified field.
Medical images like chest x-rays are of utmost importance for
the diagnosis of diseases, for instance, braintumors, dementia,
pneumonia, COVID-19, cardiovascular, and many more.
Collaborating these medical images along with the deep
learning techniques have paved the path for enormous
applications leading to the reduction in stress given upon the
health sector. In this paper, we propose a brief comparison
and study of existing technologies for analyzing chest x-rays
using deep learning.
Key Words: Chest X-Ray (CXR), Convolutional Neural
Network (CNN), Fully Convolutional Network (FCN),
Deep Cascade of Convolutional Neural Network
(DCCNN), Lookup based Convolutional neural networks
(LCNN), Deep Convolutional Neural Network(DCNN)
1. INTRODUCTION
Image segmentation is the method of dividing the images
into various pixels or regions called as segments and
processes images accurately. Diagnosis with the aid of
imaging can help in better localization and staging of the
disease. Often, Medical Images follow a traditional image
segmentation life cycle toanalyzethem.[1]Highlightsvalues
of imaging in the diagnosis of a disease. The analysisofthese
image segmentation includes pre-processing of image
followed by feature extraction and perform classification.
These processes are accomplished by using various deep
learning methodologies. The deep learning techniques from
the following researches have displayed remarkable results
for diagnosis of medical images. In the pre-processing stage,
the most commonly used technique is normalization which
helps in reducing the dimensionalityofthe imageformaking
the image fit for segmentation. Classification of images can
be done using Support Vector Machine(SVM)[3]. Otherthan
that, the data augmentation [4] is used for increasing the
robustness of deep learning. Along withthesetechniques, an
Up-sampling method [4] is implemented which helps the
network layers to propagate context data to higher
resolution layers.
Feature extraction is a technique which is used to reduce a
large input data set into relevant features. The images
required for segmentation have a large number of variables.
A lot of computing resources are required for these
variables. Therefore, feature extractions help to attain best
feature from those big data sets by selecting and integrating
variables into features, effectively decreasing the amount of
data. The final step is the classification, where we obtain the
desired output. The pre-processed and selected featuresare
provided to the model for training purposes which
ultimately gives us the appropriate result. The classification
task is a supervised learning approach in machine learning
domain. Models such as SVM in [3], U-net in [4], Resnet18,
Resnet50 and Xception in [8], LCNN and DCNN in [9], and
Fully Convolutional Network (FCN) in [15] are used for this
step in the examined research.
2. LITERATURE SURVEY
Over the past few years, the machine learning framework
has immensely helped in analyzing the complex medical
images. Tremendous advancements in medical image
segmentation anditsapplicationshavemotivatedusto study
such research techniques. The aim of [1] is to concisely
present the issue of coronavirus and highlight the
importance of medical imaging in its diagnosis. Inception
inspired, a novel deep CNN architecture where depth-wise
separable convolutions replace the Inception modules is
being discussed in [2]. Here, an architecturenamed Xception
which is similar to Inception V3 is developed which
efficiently uses the model parameters. Emphasizing
automatic tuberculosis detection [3] proposes a binary
classifier using SVM which classifies a CXR as either normal
or abnormal achieving an accuracyof82.1%andarea within
the ROC curve as 88.5%. In [4] a U-Net Network is trained

with the use of data augmentation by using elastic
deformations for effective use of annotated samples. The
segmentation results for ISBI cell tracking challenge with
average IOU (“intersection over union”)for“PhC-U373”data
set is 92% and for e “DIC-HeLa” dataset Is 77.5%. [5] In this
paper a framework for dynamic sequences of 2-D cardiac
Magnetic Resonance Images (MRI) reconstructed from
Under-sampled data using DCCNN is proposed. With the
help of this model, there is feasibility inthe reconstructionof
images with good quality through a network having
interleaved data consistency stages. It can reconstruct most
of the anatomical structures more precisely by previously
learned data. In [6] the proposed system repurposes the
networks whicharetrainedonimageclassification.Semantic
segmentation is done by using ‘Atrous Convolution’ by un-
sampled filters for feature extraction. We can explicitly
control two things: the resolution for feature response
computation and effectively amplify the field of view of
filters to integrate larger context without the need to
increase the number of parameters nor the computational
amount. In [7] thismodel showsthatconvolutional networks
when trained by themselves exceed the existing semantic
segmentation. The purpose of this research is to develop a
FCN which input the arbitrary size and createa similar-sized
output with efficient inference. This model adapts
contemporary classification networks into FCN and then
passes on their learned parameters by fine-tuning into the
segmentation block. In [8] COVID-19 and Thorax diseases,
from chest X-Ray images are being diagnosed by the system
using residual networks 18, 50 and Xception model. This
Resnet18 model gives an accuracy of 94% while resnet50
gives accuracy of 93% and the Xception model is having an
accuracy of 95%. In FUIQA [9] LCNN and DCNN are used
together for analysis wherein LCNN givesoverall accuracyof
92% and the DCNN gives overall accuracy of around 97%.
For the image quality assessment [10] CNN model is used to
reconstruct all types of images i.e., assessing the quality of
the image and also reconstructing it in a better way. The
model involves a convolutional layer, maximum and
minimum pooling followed by two fully connected layers,
and an output node. An optimization process consisting of
the network structure, feature learning and regression is
integrated, this directs to a more efficient model for judging
the image quality. In the model for image segmentation [11]
U-net is used for the segmentation of biomedical images. In
[12] for cardiovascular magnetic resonance (CMR) images,
an automated analysis method isbasedonFCN.The network
is trained and tested on an extensive dataset from the UK
Biobank. Several technical metrics,suchasthemeancontour
distance, Dice metric and Hausdorffdistance,aswell assome
medical measures, including end-systolic volume (LVESV),
LV mass (LVM); left ventricle (LV) end-diastolic volume
(LVEDV) and right ventricle (RV) end-diastolic volume
(RVEDV) and end-systolic volume (RVESV) have been used
to evaluate the system. The system gives anaccuracyof 0.93.
In [13] convolutional recurrent neural network (CRNN)
model is used for image quality assessment. High-quality
CMR images are reconstructed by CRNN. This is done with
the help of highly under sampled k-space data with
exploiting the dependencies of the temporal sequences as
well as the iterative nature of the traditional optimization
algorithms. This model gives a minimum time of 3s. In [14]
the authors have proposed the use of VGG16 for Breast
Cancer Detection giving an overall accuracy of 94.77%. For
Brain tumor analysis [15] use of deep neural networks like
FCN, Res-net, U-net, Encoder/Decoder, and CNN is done
giving an overall accuracy of 75% to 80%. Quality
assessment of Echocardiograms in [16] regression model is
enhanced to assess the standard and quality of echo images.
A stochastic gradient descent algorithm is used to evaluate
the loss function of the system. The model was designedbya
convolution stage (Convolutional and Pooling Layer) and a
fully connected stage. In the paper for image quality
assessment [17] basic operations are done to eliminate
known distortions from images. Distorted and reference
signals are scaled and aligned then metrics of digital values
need to convert into pixel intensity using nonlinear pointed
transformations. Images are decomposedintochannelsthat
are selective for spatial and temporal orientation. Discrete
Cosine Transform or Separable wavelet transform methods
are used for quality assessment methods. Errors between
decomposed references in each channel are normalized so
that presence of one image componentwill decreaseanother
image component proximate in orientation concerning the
spatial or temporal location and frequency.
Fig-1: The general steps for evaluating images using deep
learning techniques
3. RELATED WORKS
Considering the literature review in this section we will
focus upon deep learning models SVM [3], U-net [4],
Resnet18, Resnet50 and Xception [8], LCNN and CCNN [9]
and CNN [14]. In the upcoming five sections, the models will
be studied in detail and will then be compared according to
their accuracy discussed in the last section.
3.1 Abnormal CXR with tuberculosis detection
using SVM
The research presents an automated system for the
detection of tuberculosis. In [3] the performance of the
system is measured on two datasets: The MC dataset from
the Department of Health and Human Services of
Montgomery County (MC), Maryland, and the Shenzhen
dataset, from Shenzhen No.3 Hospital in Shenzhen,

Guangdong providence, China. The system uses a graph cut-
based for lung segmentation and selects training masks
according to the horizontal intensity projections of the
histogram equalized images and Bhattacharyya coefficient
for similarity projections between input and training CXR
images. Further, the subtle structures in a CXR are picked by
features computed through featurecomputationtechniques.
Object Detection Inspired Features are taken from the MC
Dataset and CBIR-based Image Features are taken from the
Shenzhen dataset where, the 594 dimensioned feature
vectors which is greater than three times the feature vector
of the MC dataset. The computation of a set of edge, shape,
and texture features is then fed as a binary classifier input.
Thus, the SVM classifier system can classify a given input
image into classified into normal or abnormal. For the MC
dataset, the classification accuracy is 78.3% and the area
under the ROC curve (AUC) is 86.9%, for the Shenzhen
dataset the classification accuracy is 82.5% and the AUC is
88%.
Model Dataset Accuracy AUC
SVM MC 78.3 86.9
Shenzhen 82.5 88
Table-1. Abnormal CXR with tuberculosis detection
3.2 U-Net: Convolution Network for segmentation
of image
The proposed architecture contains contracting paths to
capture context and symmetric expanding paths for precise
localization. This [4] model uses a weighted loss, large
weight in loss function allocated by separating labels
between touching cells, giving accurate segmentation. Data
augmentation is used for desired robust properties and
invariance when a handful of training samples are available.
The important modification in the architecture is the up-
sampling method where a large number of feature channels
are present which helps the network to propagate sensitive
information to high-resolution layers. As a result, the
expansive path and contracting path are nearly symmetric,
producing a u-shaped architecture. The proposed model
does not contain any fully connected layers and uses only
associated parts of each convolution layer which allows
smooth segmentation of arbitrarylargeimageswiththehelp
of overlap-tile strategy. In this research, the model is
evaluated using two datasets. The first data-set “PhC-U373”
contains Glioblastoma-astrocytoma U373 cells on a
polyacrylamide substrate recorded by phase-contrast
microscopy. Here, the model achieves an average IOU
(Intersection Over Union) of 92%, which is significantly
better than the prior best algorithm (Sliding-window
Convolutional Network) with an IOU of 83%. The second
data-set “DIC-HeLa” are HeLa cells ona flatglassrecordedby
differential interference contrast(DIC)microscopy.Herethe
achieved IOU average of 77.5% which is significant in front
of the prior best algorithm with 46%.
Model Dataset IOU
U-Net Phc-U373 92%
DIC-HeLa 77.5%
Table-2. U-Net based image segmentation
3.3 Image Quality Assessment for fetal ultrasound
with Deep Convolutional Neural Networks
This paper [8] proposes the use of LCNN and CCNN for
classification purposes for the image quality assessment for
fetal ultrasounds. The performance of the system is
evaluated based on data taken from Shenzhen Maternal and
Child Healthcare Hospital from 6 September 2012 to
November 2013. The range of fetal ages were between 16 to
40 weeks. In the proposed paper, uses two deep
convolutional neural networksnamelyLCNN andCCNN with
an accuracy of 96%. The L-CNN purpose is to find the region
of interest (ROI) of the fetal abdominal in the image.
According to the ROI discovered by the L-CNN, the C-CNN
assesses the quality of an image by determiningitsgoodness
of deciding the essential structures of the stomach bubble
and umbilical vein. Thereafter, for increasing the
performance of the L-CNN, augmentation of data is done
between the input of the neural network containing the
features of local phase with the original data collected. The
experimental evaluations of the scheme are carried out in
three parts. In the first part, the L-CNN and C-CNN
implementation issues are investigated. In the second part,
the performance of the C-CNN model before and after
training with visualization of the model is illustrated
quantitatively. To illustrate the performance of the C-CNN
and LCNN, a comparison of the results of the region of
interest (ROI) identification and the assessment of SB
(stomach bubble) and UV (umbilical vein) structures from
FUIQA scheme is done.Assessmentmetricssuchasaccuracy,
sensitivity and specificity are selected for comparing
quantitatively. In this method, the outcomes according to 3
doctors E1 E2 and E3 are considered
Doctors ROI SB UV
Accuracy Accuracy Accuracy
E1 0.93 0.99 0.98
E2 0.91 0.99 0.97
E3 0.92 0.98 0.98
Table-3. Fetal ultrasound image quality assessment

3.4 ImageSegment-basedClassificationforChestX-
Ray
The system [9] uses U-net with Xception encoder and
Feature Pyramid network with Xception encoder and
Resnet50 encoder for segmentation of images and Xception,
Resnet50, ResNet18 for classification purposes with an
accuracy of 94%, 93%, and 95% respectively. This system is
based on Deep Learning and AI Summer/Winter School
(DLAI) Hackathon Phase 3—Multiclass COVID-19 Chest X-
ray Challenge dataset and was collected from various
sources. The dataset comprises of three classes that are
“COVID-19”, “Thorax-Diseases”, and “Clear”. Inordertotune
with the CNN model, the original RGB X-Ray images were
converted into monochrome images along with floating-
point pixels. The dataset was randomly split intotrainingset
and test sets with an 80–20 split. Also, augmentation
increased the number of the training set and enabled the
model to become robust and aids in a better comprehension
uncommon CXR images. A random selection of one or more
of the following operations: random cropping of CXR with
concentration on lung area, horizontal and vertical flipping,
90-degree rotation, and histogram equation were
implemented on the original image.Tosegmentthelungpart
a deep learning model for semantic segmentation was used.
A CNN model with ResNet18, Resnet 50, and Xception was
used. ResNet is residual learning model which skips
connections to resolve the gradientvanishingproblemalong
the increasing depth of CNN layers.Theoverall accuracywas
enhanced by Xception: A CNN model, established upon
depth-wise separable convolution layers. The cross-channel
mapping and spatial correlations in the feature mapsofCNN
can be entirely decoupled and was linearly assembled with
ResNet connections.
Method Cross Entropy
Accuracy Precision Recall F1
ResNet18 94 90 89 89
ResNet50 93 87 90 88
Xception 95 94 90 92
Average 94 90.3 89.7 89.7
Table-4.a. Chest X-Ray Classification(Cross Entropy)
Method Focal Loss
Accuracy Precision Recall F1
ResNet18 94 91 91 91
ResNet50 93 87 90 89
Xception 94 91 93 92
Average 93.7 89.7 91.3 90.7
Table-4.b. Chest X-Ray Classification(Focal Loss)
3.5 Breast Cancer Medical Image Classification
using CNN
In this system [14], CNN Model is used primarily. Model pre-
processes the images with specified parameters of medical
images required to fit in the network. Imagesarein224x224
pixel PNG format with rescaled boundaries of images. For
normalizing the images, reducing complexity of topology
function and increasing optimization of learning of model,
implementation of new level of orthogonality between the
layer is introduced. GPU performseliminationofover-fitting,
zooming, and translation with zero computation cost. For
given 15 layered CNN model, input used is VGG16
conventional specified images with dimension 224x224x3
with RGB color channels. There is batch size of 8 is used due
to GPU requirement. The model will have epoch set to 100.
The preferred optimizer used in work is stochastic gradient
descent (SGD) because of faster and frequent updates. The
activation function used is ReLU as having good
computational efficiency and ability to work in over fitting
case. The class mode used is Binary i.e., Malignant and
benign. Learning rate being 0.1 and momentum set to 0.9.
Strides fixed t 1 pixel. The convolutional layer has 64 filters
of 3x3 dimension outputting 224x224x64. The fully
connected layer has 2 class output and dropout probability
up to 50% to work properly with overfitting in the model.
Model Training
Accuracy
(%)
Val.
Accuracy
(%)
Training
Loss
(%)
Val.
Loss
(%)
VGG16 94.77 96.33 13.94 10.90
CNN 98.02 98.50 5.84 4.81
Table-5. Brest cancer medical image classification
4. PERFORMANCE COMPARISION
For biomedical images segmentation, the deep learning
approach surpasses the accuracy. In section E, the CNN
model outperforms other models in terms of performance.
The combination of CNN classification and image
segmentation produces a finegrained detection model. The
results show that training for multiple iterations on a
designed model, accuracy is obtained are best among all.
Implementation of a new level of orthogonality between the
layers is introduced to increase learning in the process. The
use of zero computational cost GPU for image augmentation
purposes. In sections A and B we spot that the accuracy of
models varies as the dataset differs. For section A, SVM has
been applied to two datasets (MC and Shenzen), it is seen
that from Table 1 the accuracy for the dataset Shenzen is
more than that of MC. Similarly in section B, U-net has
worked on two different datasets. From Table2 we
acknowledge that the accuracy for the dataset PhC-U373 is
greater than that for DIC-HeLA. There are multiple reasons
for the change in accuracy on different datasets irrespective

of the same technique applied. In section D, Xception
encoder which is a CNN model, using depth-wise separable
convolutional layers with cross channel mapping, feature
maps of CNN can be entirely uncoupled and were stacked
linearly with residual connections causing an increaseinthe
accuracy. In Section C, LCNN and CCNN for image
classification of fatal ultrasound image quality. According to
the ROI discovered by the L-CNN, the C-CNN evaluates the
quality of an image by determining its goodness of depiction
for the key structures of the image data. Experiments
conducted using different deep learning models, Xception,
LCNN+CCNN, Resnet18 and Resnet50achieveperforms well
after CNN. Fine-tuning avoiding losses give higher accuracy
when used on a larger dataset.
5. EXPERIMENTAL RESULTS
Fig-2. Models according to increasing accuracy
From this study, we have deduced that the 15-layer CNN
model most effectively classifies and analyzes the medical
images using relevant CNN classification techniques.
Xception model along with the CNN model gives the second-
best accuracy. Further, we propose the implementation of
CNN models with different activation functions and focuson
removing the noise for better segmentation of the images
hence improving the performance.
6. CONCLUSIONS
Medical Image analysis is mainly used for the detection and
diagnosis of diseases which helps in proper and accurate
treatment given to patients. In this paper we have taken an
overview of various deep learning models focusing on the
classification task. This study has concluded that a union of
traditional analysis approaches along with deep learning
framework upshots in an effective diagnosis of brain tumor,
fatal ultrasound, COVID 19, thorax, etc. Amongst the
following models U-net, SVM, Residual networks, CNN, and
Xception, the highest accuracy is achieved bytheCNN model
of 98.0.2%. The study illuminates that the drawbacks of a
single model are overcome by encouraging a combination of
deep learning models. In future works, this study can assist
in enhancing the system using advanced deep learning
frameworks.
REFERENCES
[1] Wenjing Yang, Arlene Sirajuddin, Xiaochun Zhang,
Guanshu Liu, Zhongzhao Teng, Shihua Zhao, and Minjie Lu.
2020. The role of imaging in 2019 novel coronavirus
pneumonia (COVID-19). European Radiology 30, 9 (April
2020), 4874–4882. https://doi.org/10.1007/s00330-020-
06827-4.
[2] F. Chollet. 2017. Xception: Deep Learning with
Depthwise Separable Convolutions. In Proceedings of the
IEEE Conference on Computer Vision and Pattern
Recognition (CVPR 2017). IEEE, Honolulu, HI, 1800–1807.
https://doi.org/10.1109/CVPR.2017.195.
[3] S. Jaeger, A. Karargyris, S. Candemir, L.Folio,J.Siegelman,
F. Callaghan, Z.Xue, K. Palaniappan, R. K. Singh, S. Antani, G.
Thoma, Y. Wang, P. Lu, and C. J. McDonald. 2014. Automatic
Tuberculosis Screening Using Chest Radiographs. IEEE
Transactions on Medical Imaging 33,2(Feb2014),233–245.
[4] Olaf Ronneberger, Philipp Fischer, and Thomas Brox.
2015. U-net: Convolutional networks for biomedical image
segmentation. In Proceedings of the International
Conference on Medical Image Computing and Computer-
Assisted Intervention (MICCAI 2015). Springer, Munich,
Germany, 234–241. https://doi.org/10.1007/978-3-319-
24574-4_28
[5] J. Schlemper, J. Caballero, J. V. Hajnal, A. N. Price, and D.
Rueckert, “A deep cascade of convolutional neural networks
for dynamic MR image reconstruction,”IEEEtransactions on
Medical Imaging, vol. 37, no. 2, pp. 491– 503, 2017.
[6] Liang-ChiehChen,George Papandreou,IasonasKokkinos,
Kevin Murphy, and Alan L Yuille. 2018. Deeplab: Semantic
image segmentation with deep convolutional nets, atrous
convolution, and fully connected crfs. IEEE transactions on
pattern analysis and machine intelligence40,4(April 2018),
834–848.
[7] J. Long, E. Shelhamer, and T. Darrell, “Fully convolutional
networks for semantic segmentation,”inProc.CVPR,2015,
pp. 3431–3440.
[8] Phongsathorn Kittiworapanya, Kitsuchart Pasupa, “An
Image Segmentbased Classification for ChestXRayImage” in

ACM November 19, 2020, https://doi.org/10.1145/
3429210.3429227.
[9] L. Wu, J.-Z. Cheng, S. Li, B. Lei, T. Wang, and D. Ni, “FUIQA:
Fetal ultrasound image quality assessment with deep
convolutional networks,” IEEE Transactions onCybernetics,
vol. 47, no. 5, pp. 1336–1349, 2017.
[10] L. Kang, P. Ye, Y. Li, and D. Doermann, “Convolutional
neural networks forno-referenceimage qualityassessment,”
in Proceedings of the IEEE conference on computer vision
and pattern recognition, 2014, pp. 1733– 1740
[11] O. Ronneberger, P. Fischer, and T. Brox, “U-net:
Convolutional networksforbiomedical imagesegmentation”
in MICCAI. Springer, 2015, pp. 234–241.
[12] W. Bai, M. Sinclair, G. Tarroni, O. Oktay, M. Rajchl, G.
Vaillant, A. M. Lee, N. Aung, E. Lukaschuk, M. M. Sanghvi,
“Automated cardiovascular magnetic resonance image
analysis with fully convolutional networks” in Journal of
Cardiovascular Magnetic Resonance, vol. 20, no. 1, p. 65,
2018.
[13] C. Qin, J. Schlemper, J. Caballero, A. N. Price, J. V. Hajnal,
and D. Rueckert, “Convolutional recurrentneural networks
for dynamic MR image reconstruction,”IEEEtransactions on
medical imaging, vol. 38, no. 1, pp. 280–290, 2018.
[14] Yongbin Yu, Ekong Favour, Pinaki Mazumder,
Convolutional Neural Network Design for Breast Cancer
Medical Image Classification , 2020 IEEE 20thInternational
Conference on Communication Technology
[15] MAHNOOR ALI, SYED OMER GILANI, ASIM WARIS,
KASHAN ZAFAR, AND MOHSIN JAMIL, BrainTumourImage
Segmentation Using Deep Networks IEEE Open Access
Journal Volume 8, 2020.
[16] A. H. Abdi, C. Luong, T. Tsang, G. Allan, S. Nouranian, J.
Jue, D. Hawley, S. Fleming, K. Gin, J. Swift, et al., “Automatic
quality assessment of apical four-chamberechocardiograms
using deep convolutional neural networks,” in Proc. SPIE,
vol. 10133, 2017, pp. 101 330S– 1.
[17] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli,
“Image quality assessment: From error visibility to
structural similarity,” IEEE Trans.Image Process.,vol.13, no.
4, pp. 600– 612, Apr. 2004.

Performance Comparison Analysis for Medical Images Using Deep Learning Approaches

More Related Content

Similar to Performance Comparison Analysis for Medical Images Using Deep Learning Approaches (20)

More from IRJET Journal (20)

Recently uploaded (20)

Performance Comparison Analysis for Medical Images Using Deep Learning Approaches