Performance evaluation of different automatic seed point generation techniques for segmentation of indian vehicle number plate

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976-6367(Print),
ISSN 0976 - 6375(Online), Volume 6, Issue 5, May (2015), pp. 36-46© IAEME
36
PERFORMANCE EVALUATION OF DIFFERENT
AUTOMATIC SEED POINT GENERATION TECHNIQUES
FOR SEGMENTATION OF INDIAN VEHICLE NUMBER
PLATE
Veena M.N1
, Arpitha K.S2
, Vasudev T3
1
P.E.T Research Foundation, P.E.S. College of Engineering, Mandya, Karnataka, India
2,3
Maharaja Institute of Technology, Mysore, Karnataka, India
ABSTRACT
The Vehicle Number Plate Recognition (VNPR) system plays an important role in traffic
surveillance systems, such as traffic law enforcement, real time monitoring, parking systems, road
monitoring and security systems. The detection and extraction of number plate system is a real time
embedded module in which automatically localize the vehicle number plate for further process of
reading. The proposed work focus on performance evaluation for different key seed points
generation techniques applied for number plate segmentation from vehicle images. Harris corner
technique along with combination of region based and threshold technique shows better performance
results in terms of computational efficiency and better segmentation. A detailed experimentation is
carried out for performance evaluation.
Keywords: Harris corner detector, Segmentation of number plate, Seed Point generation, Threshold.
1. INTRODUCTION
Now a days every country is allocating sufficient budget on traffic automation and vehicle
theft controlling. This is because of the enormous increase in the vehicles on roads and this leads to
limitations/weaknesses such as leak checking, false checking. In order to avoid such limitations an
intelligent system for the traffic surveillance is required replace the traditional way of finding the
theft vehicle or traffic rules violated vehicles in its own way.
The most vital and the difficult part in the system is segmentation of the vehicle number
plate, which directly affects the overall efficiency of the system. The presence of noise, blurring in
INTERNATIONAL JOURNAL OF COMPUTER ENGINEERING &
TECHNOLOGY (IJCET)
ISSN 0976 – 6367(Print)
ISSN 0976 – 6375(Online)
Volume 6, Issue 5, May (2015), pp. 36-46
© IAEME: www.iaeme.com/IJCET.asp
Journal Impact Factor (2015): 8.9958 (Calculated by GISI)
www.jifactor.com
IJCET
© I A E M E

37
the image, uneven illumination, dirt, rain, dim light and foggy conditions make the task even more
complex.
Three major phases involved in localization of Indian Vehicle number plate are automatic
seed points generation, segmentation and extraction. In first phase, automatic seed points are
generated using different techniques such as Harris corner, Harris Laplace, Laplacian of the
Gaussian, Gilles and Susan. In second phase, seed points generated by each technique which is in
specific range is taken as a initial seed point and region starts to grow from initial seed. Growth of
region depends upon the intensity value of the neighboring pixels as well as threshold value. The
difference between a pixel's intensity value and the region's mean, is used as a measure of similarity.
The pixel with the smallest difference measured this way is allocated to the region. This process
stops when the intensity difference between region mean and new pixel becomes larger than a certain
threshold. Finally the resulting region will be the required segmentation region [1-3]. In third phase,
extraction of the segmented candidate number plate region is done by using bounding box method.
2. LITERATURE SURVEY
Literature survey indicates that quite a number of researchers have explored useful methods
for locating number plate from the vehicle images. Some of the methods are based on normal
features of number plates like color [4], shape [5], symmetry [6], texture of grayness [7], spatial
frequency [8] and variance of intensity values [9]. An approach is suggested based on enhanced
detection of boundary lines [10] using gradient filter. Another similar attempt for detection of
bounding lines based on two pair of parallel lines using Hough transform [11] to designate number
plate is reported. Other approaches are reported based on the morphology of objects in an image
[12,13]. These approaches focus on some salient properties of vehicle plates such as their brightness,
contrast, symmetry, angles etc, in an image and locate the position of number plate regions. Two
approaches are based on statistical properties of text [14,15]. In these approaches, text regions are
discovered using statistical properties of text like the variance of gray level, number of edges, edge
densities in the region etc. These approaches were commonly used in finding text in images and used
to detect and extract number plate areas as they contain alphabets and numerals.
In addition, vehicle number plate detection using artificial intelligence (AI) and genetic
algorithm [16, 17] are proposed. These systems use edge detection and edge statistics and then AI
techniques are applied to detect the location of the number plate area. The works reported in
literature have some kind of limitations such as plate size dependency, color dependency, efficient
only in certain conditions or environment like indoor images etc.
The work proposed in [18] is an algorithm using novel adaptive image segmentation
technique (sliding concentric windows) for vehicle license plate identification. In this paper, number
plate segmentation has been addressed through the implementation of SCW segmentation method,
image masking, Binarization with Sauvola method, and finally connected component labeling and
binary measurements, which are arranged in sequence. The SCW deals with the detection of the
region of interest. Connected Components Analysis (CCA) [19] is a well known technique in image
processing that scans an image and labels its pixels into components based on pixel connectivity.
Once all groups have been determined, each pixel is labeled with a value according to the component
it was assigned to. The input images taken are gray scale images and the method claims 96.5%. of
efficiency in segmentation.
Another work [20] was proposed using a wavelet transform method for the extraction of
important contrast features as guides in searching for number plate. In the wavelet transform, there
are four subimages (subbands), namely LL, LH, HL, and HH, where L and H stand for low and high
frequency, respectively. According to this paper, a reference line in the first-level LH subband (1LH)
subimage exactly above the plate is noticeable. Using the above reference line, a searching mask is

International Journal of Computer Engineering and Technology (IJCET), ISSN 0976
ISSN 0976 - 6375(Online), Volume 6, Issue
created, speeding up the execution time. The average accuracy of de
Nevertheless, the method is unreliable when the distance between the vehicle and the acquisition
camera is either too far or too close or the angle of viewpoint is wide.
Masood et al. detected corners for planar curves by sliding s
curve and counting number of contour po
boundary-based corner detection method using wavelet transform for its ability for detecting sharp
variations [22, 23].
3. PROPOSED MODEL
The proposed model considered for performance evaluation
different seed point generation technique for number pla
this system is vehicle image that is acquired thro
illumination conditions and different background
3.1 Automatic Seed point Generation by Harris Corner Detector
Harris corner detection theory [
rotation, scale, illumination variation and image noise. In Harris corner detector method the local
autocorrelation function measures the local changes of the signal with patches shifted and the
discreteness refers to the shifting of the patches. Given a shift
correlation function is given in eq.(1)
6375(Online), Volume 6, Issue 5, May (2015), pp. 36-46© IAEME
38
created, speeding up the execution time. The average accuracy of de
camera is either too far or too close or the angle of viewpoint is wide.
Masood et al. detected corners for planar curves by sliding set of three rectangles along the
curve and counting number of contour points lying in each rectangle [21]. Peng et al. introduced a
based corner detection method using wavelet transform for its ability for detecting sharp
proposed model considered for performance evaluation for number plate extraction
different seed point generation technique for number plate segmentation shown in Fig.3.
this system is vehicle image that is acquired through digital camera. Images are acquired in differen
different background.
Fig 3.1 Proposed Model
Automatic Seed point Generation by Harris Corner Detector
Harris corner detection theory [24] used to realize the seed points selection is invariant to
fting of the patches. Given a shift (δx,δy) and a point
q.(1) :
© IAEME
created, speeding up the execution time. The average accuracy of detection was 92.4%.
et of three rectangles along the
]. Peng et al. introduced a
based corner detection method using wavelet transform for its ability for detecting sharp
for number plate extraction using
te segmentation shown in Fig.3.1. Input to
Images are acquired in different
ints selection is invariant to
and a point (x,y), the auto

39
, = ∑ , y [I x + δx, y + δy − I x , y ] (1)
where I(x,y) denotes the image function and (x , y ) are the points in the Gaussian window W
centered on (x,y). The window function W centered on (x,y) is given in eq.(2) :
W x, y = πσ
e / σ
(2)
The shifted image is approximated by a Taylor expansion truncated to the first order terms is given
in eq.(3) :
I x +δx, y +δy)=I x ,y )+[I I ]!
δ
δ
" (3)
where #$ and # denote the partial derivatives in x and y respectively. Substituting eq.(3) into eq.(1)
is given in eq.(4) :
S(x,y)=(δx,δy)S(x,y)!
δ
δ
" (4)
where S(x,y) captures the intensity structure of the local neighborhood. Let ʎ , ʎ be the Eigen
values of S(x,y) then the values obtained for ʎ and ʎ directs as ∶ if ʎ , ʎ are small then the local
auto-correlation function is flat, the windowed image region is of approximately constant intensity.
In ʎ , ʎ if one is high and other is small then local auto-correlation function is ridge shaped and
this indicates an edge. If both ʎ , ʎ values are high then local auto-correlation function is sharply
peaked, and this indicates a corner. The seed points generation from this method is illustrated in Fig
3.1.1 through 3.1.4.
Fig 3.1.1 Input Vehicle Image Fig. 3.1.2 Automatic seed points Fig.3.1.3 Segmented Number
Generation by Harris corner plate region
Method
Fig 3.1.4 Extracted Number Plate
This method gives ideal segmentation result using Harris Corner detector.

40
3.2. Automatic Seed point Generation by Laplacian of Gaussian (LoG) Detector
Lindeberg proposed a detector[25] for blob like features that searches for scale space
extrema of a scale-normalized Laplacian of Gaussian(LoG) is given in eq.(5):
L(x,σ)=σ (I (x,σ)+I (x,σ)) (5)
Thus, given a discrete two-dimensional input image f(x,y) a three dimensional discrete scale-
space L(x,y,σ) is computed and a point is regarded as a bright (dark) blob if the value at this point is
greater than the value in all its 26 neighbours.
The LoG filter mask corresponds to a circular center-surround structure, with positive
weights in the center region and negative weights in the surrounding ring structure. Thus, it will
yield maximal responses if applied to an image neighborhood that contains a similar (roughly
circular) blob structure at a corresponding scale. By searching for scale space extrema of the LoG,
we can therefore detect circular blob structures. Note that for such blobs, a repeatable key point
location can also be defined as the blob center. The LoG can thus both be applied for finding the
characteristic scale for a given image location and for directly detecting scale-invariant regions by
searching for 3D (location + scale) extreme of the LoG. The seed points generation from this method
is illustrated in Fig 3.2.1 through 3.2.4
Fig 3.2.1 Input Vehicle Image Fig. 3.2.2 Automatic seed point Fig.3.2.3 Segmented Number
generated by LOG detector plate Region
Fig 3.2.4.Extracted Number Plate
This method gives over segmentation result using LoG detector
3.3. Automatic Seed point Generation by Gilles method
Gilles proposed an information-theoretic algorithm [26] in which seed points correspond to
image locations at which the entropy of local intensity values attains a maximum. Gilles used the
Shannon entropy of local attributes such as the intensity of the pixels in a neighborhood around the
current pixel. Image areas with a flatter distribution of pixel intensities have a higher signal

41
complexity and thus a higher entropy. In contrast, a flat image region has a peaked distribution. As
Gilles used a fixed size neighborhood, the algorithm selected only those salient points which were
appropriate to the size of the neighborhood. As an extension, Gilles proposed the use of a global
scale for the entire image, which was automatically selected by searching for peaks in the average
global saliency for increasing scales. The seed points generation from this method is illustrated in
Fig 3.3.1 through 3.3.4.
Fig 3.3.1 Input Vehicle Image Fig. 3.3.2 Automatic seed point Fig.3.3.3 Segmented Number
generation using Gilles method plate Region
Fig 3.3.4 Extracted Number
This method gives average segmentation result using Gilles method
3.4. Automatic Seed point Generation by Harris Laplace method
The Harris-Laplacian operator [25] was proposed for increased discriminative power
compared to the Laplacian. It combines the Harris operator’s specificity for corner-like structures
with the scale selection mechanism by eq(5). The method first builds up two separate scale spaces
for the Harris function and the Laplacian. It then uses the Harris function to localize candidate points
on each scale level and selects those points for which the Laplacian simultaneously attains an
extremum over scales. The resulting points are robust to changes in scale, rotation, illumination, and
camera noise. As a drawback, however, the original Harris-Laplacian detector typically returns a
much smaller number of points than the Laplacian or DoG detectors. The seed points generation
from this method is illustrated in Fig 3.4.1 through 3.4.4

Fig 3.4.1 Input Vehicle Image Fig. 3.4.2 Automatic seed point
Fig 3.
This method gives better segmentation result using LoG detector
3.5. Automatic Seed point Generation by Susan corner method
SUSAN [27] is an acronym standing for
feature detection, SUSAN places a circular mask over the pixel to be tested (the nucleus). The region
of the mask is , and a pixel in this mask is represented by
pixel is compared to the nucleus using the c
where determines the radius,
has been determined empirically. This
rectangular function. The area of the SUSAN is given by
If is the rectangular function, then
nucleus. The response of the SUSAN operator is given by
42
put Vehicle Image Fig. 3.4.2 Automatic seed point Fig.3.4.3 Segmented Number
Generation by Harris Laplace Plate Region
Method
Fig 3.4.4Extracted Number Plate
This method gives better segmentation result using LoG detector
. Automatic Seed point Generation by Susan corner method
is an acronym standing for smallest univalue segment assimil
, and a pixel in this mask is represented by . The nucleus is at
s using the comparison function is given by eq (6)
(6)
determines the radius, is the brightness of the pixel and the power of the exponent
has been determined empirically. This function has the appearance of a smoothed
. The area of the SUSAN is given by eq (7):
(7)
is the rectangular function, then is the number of pixels in the mask which are within
nucleus. The response of the SUSAN operator is given by eq (8):
(8)
© IAEME
Fig.3.4.3 Segmented Number
Harris Laplace Plate Region
smallest univalue segment assimilating nucleus. For
. The nucleus is at . Every
eq (6):
is the brightness of the pixel and the power of the exponent
function has the appearance of a smoothed top-hat or
is the number of pixels in the mask which are within of the

where is named the `geometric th
score if the area is small enough. The smallest SUSAN locally can be found using non
suppression, and this is the complete SUSAN operator.
The value determines how similar points have
considered to be part of the univalue segment. The value of
univalue segment. If is large enough, then this becomes an
further steps are used. Firstly, the
centroid far from the nucleus. The second step insists that all points on the line from the nucleus
through the centroid out to the edge of the mask are in the SUSAN
this method is illustrated in Fig 3.5.1 through 3.5.
Fig 3.5.1 Input Vehicle Image
Fig 3.5.4
This method gives over segmentation result using Susan Corner Detetor.
4. EXPERIMENTAL RESULTS
The automatic seed points region based segmentation methods are implemented using
MATLAB R2011b on Intel(R) Core 2 Duo processor @ 2.20 GHz and
vehicle images are considered for experimentation. The experimental results for computational
performance evaluation conducted for different seed points generation methods for number plate
segmentation from vehicle images are tabulated
Table.1 Results of computational efficiency of different seed points
Automatic Seed points
generation
Methods
Harris corner
Gilles
LoG
Harris Laplace
Susan corner
43
is named the `geometric threshold'. In other words the SUSAN operator only has a positive
score if the area is small enough. The smallest SUSAN locally can be found using non
suppression, and this is the complete SUSAN operator.
determines how similar points have to be to the nucleus before they are
considered to be part of the univalue segment. The value of determines the minimum size of the
is large enough, then this becomes an detector. For
rstly, the centroid of the SUSAN is found. A proper corner will have the
oid out to the edge of the mask are in the SUSAN. The seed points generation from
3.5.1 through 3.5.4
Fig. 3.5.2 Automatic seed point Fig.3.5.3 Segmented Number
generation using Susan corners
detector
Fig 3.5.4 Extracted Number Plate
This method gives over segmentation result using Susan Corner Detetor.
4. EXPERIMENTAL RESULTS
MATLAB R2011b on Intel(R) Core 2 Duo processor @ 2.20 GHz and RAM 2 GB. About 120
segmentation from vehicle images are tabulated in Table.1.
computational efficiency of different seed points generation techniques
Average run time to generate
seed points (seconds)
Average run time to segment
vehicle number plate (seconds)
0.69
8.8
8.2
147.6
13.75
© IAEME
reshold'. In other words the SUSAN operator only has a positive
score if the area is small enough. The smallest SUSAN locally can be found using non-maximal
to be to the nucleus before they are
determines the minimum size of the
detector. For corner detection, two
of the SUSAN is found. A proper corner will have the
The seed points generation from
.3 Segmented Number
plate Region
RAM 2 GB. About 120
generation techniques
run time to segment
vehicle number plate (seconds)
8.69
22.49
24.7
10.86
12.13

44
The average results observed are: Harris corner takes 0.69 secs, LoG detector takes 8.2 secs,
Gilles takes 8.8 secs, Harris Laplace takes 147.6 secs and susan corner takes 13.75 secs respectively
to compute seed points. The average run time to segment vehicle number plate using region based
segmentation along with each automatic seed point generation is Harris corner takes 8.69 secs, LoG
detector takes 24.7 secs, Gilles takes 22.49 secs, Harris Laplace takes 10.86 secs and susan corner
takes 12.13 secs respectively. The performance evaluation results are shown in Fig 4.1.
Fig 4.1 Comparison of computational efficiency of different seed points generation techniques
The correct segmentation rate observed for Harris Corner is 67%, Harris Laplace is
60%,Susan corner is 21%,LoG is 17% and Gilles is 24% and incorrect segmentation rate is Harris
Corner is 33%, Harris Laplace is 40%,Susan corner is 79%,LoG is 83% and Gilles is 76%
respectively. Table 2 shows the results of vehicle number plate segmentation using different seed
points generation techniques.
Table.2 Results of Vehicle Number plate Segmentation
5. CONCLUSION
In the proposed work a comparative analysis of finding better seed points generation
techniques for number plate segmentation is performed. The evaluation analysis detects the Harris
corner detector is the best suited method. Since it is invariant to rotation, scale, illumination variation
and image noise, more over the method supports the segmentation relatively better than the other
approaches. The method finds local autocorrelation to measure the local changes of the signal with
patches shifted and uses computationally less time. The region based segmentation using Harris
corner detector gives 67% success and failure due to lack of generating initial seed point in order to
Automatic seed
points
generation
methods
Number of Input
Vehicle Images
Number of
successfully
segmented Number
plates
Success rate
(%)
Failure rate (%)
Harris corner 120 80 67 33
Harris Laplace 120 72 60 40
Susan corner 120 25 21 79
LoG 120 20 17 83
Gilles 120 29 24 76
0.69
8.8 8.2
147.6
13.758.69
22.49 24.7
10.86 12.13
0
20
40
60
80
100
120
140
160
Harris corner Gilles LoG Harris Laplace Susan corner
Averageruntimetakenin
seconds
Automatic seed points generation techniques

45
produce region of interest, when vehicle body and number plate are of same color or similar
intensity and also if vehicle number plate covered by shadow. In the future work it is planned to
take care to overcome the above failure cases.
REFERENCES
1. Prince Pal Singh, Jaswinder Singh, “Automatic Seed Placement in Region Growing Image
Segmentation” Journal of Engineering, Computers & Applied Sciences (JEC&AS), Volume
2, No.7, July 2013.
2. S Jayaraman, S Esakkirajan, T Veerakumar “Digital Image Processing”, 2011.
3. Verma, O.P. Dept. of Inf. Technol., Delhi Technol. Univ., Delhi, India Hanmandlu, M. ;
Susan, S. ; Kulkarni, M. ; Jain, P.K “A Simple Single Seeded Region Growing Algorithm for
Color Image Segmentation using Adaptive Thresholding”, 2011.
4. Zhu Wei-gang H.G-j, Jia Xing, "A study of locating vehicle license plate based on color
feature and mathematical morphology ", In the 6th International Conference on Signal
processing, Beijing, 2002, vol.1, pp. 748-751.
5. Ahmed, M.Sarfaz, A. Zidouri and K.G AI-Khatib, " License plate recognition system", In the
10th IEEE International Conference on Electronics, Circuits and Systems, 2003, pp.898-901.
6. Kim D.S and S.I Chien, "Automatic car license plate extraction using modified generalized
symmetry transform and image wrapping ", In the IEEE Symposium on Industrial
Electronics, 2001, vol.3,pp.2022-2027.
7. Christos Nikolaos .E.Anagnostopoulous, I.E.A and V.L.A.E Kayafas, " A License Plate
recognition algorithm for Intelligent transportation System Applicatios",IEEE Transactions
on Intelligent Transportation Systems, 2006,7(3), pp.377-392 .
8. Hsieh C.Y, Y.S Juan and K.M Hung "Multiple license plate detection for complex
background", In the 19th IEEE International Conference on Advanced Information
Networking and Applications, 2005, vol.2, pp.389-392.
9. Gao D.S. and Zhou, "Car License Plates detection from complex scene ", In the 5th
International Conference on Signal Processing, 2000, vol2. pp.1409-1414.
10. Duan T.D, D.A.Duc, T.L.H Due , "Combining Hough Transform and Contour Algorithm for
detecting Vehicles License Plates", Proceedings of International Symposium on Intelligent
Multimedia, Video and Speech Processing , 2004, pp.747-750.
11. Remus.B. " License Plate Recognition System", Proceedings of the 3rd International
Conference in Information, Communications and Signal Processing, 2001,pp.203-206
12. Bai H.L, C.P.Liu," A Hybrid License Plate Extraction Method Based on Edge Statistics and
Morphology", Proceeding of the 17th International Conference on Pattern Recognition 2004
13. Gonzalez .R.C, R.E .Woods ,"Digital Image Processing", 2nd Edition, Printice Hall,
Englewood Cliffs , NY 2002.
14. Clark .P.M.Mirmehdi, "Finding Text Regions using Localised Measures", Proceedings of the
11th British Machine Vision Conference, 2000, pp.675-684.
15. Clark P.M. Mirmehdi, " Combining Statistical Measures to find Image Text Regions ",
Proceedings of the 15th International Conference on Pattern Recognition, 2000 pp.450-453
16. Bishop C.M, "Neural Networks for Pattern Recognition" Oxford: Clarenden Press 1995
17. Parisi.R, E.D, Di Claudio, G.Lucarelli, G.Orlandi, "Car Plate Recognition by Neural
Networks and Image Processing" Proceedings of the 1998 IEEE International Symposium on
Circuits and Systems.1998, pp. 195-198.
18. C. Anagnostopoulos, I. Anagnostopoulos, E. Kayafas, and V. Loumos, “A license plate
recognition system for intelligent transportation system applications,” IEEE Trans. Intell.
Transp. Syst., vol. 7, no. 3, pp. 377–392, Sep. 2006.

46
19. S.E.Umbaugh, Computer Vision and Image Processing, Prentice Hall, New Jersey, 1998,
pp.133-138
20. C.-T. Hsieh, Y.-S. Juan, and K.-M. Hung, “Multiple license plate detection for complex
background,” in Proc. Int Conf. AINA, 2005, vol. 2, pp. 389–392.
21. Asif Masood, M. Sarfraz. “Corner detection by sliding rectangles along planar curves,”
Computers & Graphics,Vol. 31, pp.440-448, 2007.
22. X. Peng, C. Zhou, M. Ding, “Corner detection method based on wavelet transform,” In: Proc.
SPIE, Vol. 4550, pp.319-323, 2001.
23. G. Kontogianni, E. K. Stathopoulou*, A. Georgopoulos, A. Doulamis “HDR Imaging for
feature detection on detailed architectural designs” The international archives of the
photogrammetry, remote sensing and spatial information science,volumeXL-5/WA,2015.
24. D. Muhammad Noorul Mubarak, M. Mohamed Sathik, S.Zulaikha Beevi and K. Revathy ,” A
hybrid region growing algorithm for medical Image segmentation,” International Journal of
Computer Science & Information Technology (IJCSIT) Vol 4, No 3, June 2012.
25. Kristen Grauman and Bastian Leibe "Local Features: Detection and Description"
26. S. Gilles, Robust matching and description of images. PhD thesis, Oxford University, Oxford,
U.K., 1998.
27. Jie Chen, Li-hui Zou, Juan Zhang and Li-hua Dou “The Comparison and Application of
Corner Detection Algorithms” Journal of Multimedia, vol. 4, NO. 6, December 2009.
28. M. M. Kodabagi and Mr. Vijayamahantesh S. Kanavi, “License Plate Recognition System
For Indian Vehicles” International Journal of Advanced Research in Engineering &
Technology (IJARET), Volume 4, Issue 2, 2013, pp. 295 - 304, ISSN Print: 0976-6480, ISSN
Online: 0976-6499.
29. Prof. D. N. Rewadkar and Tuhina Dixit, “Least-Congested Route Estimation Using GPS
Equipped Vehicles In Urban Road Networks” International journal of Computer Engineering
& Technology (IJCET), Volume 5, Issue 4, 2014, pp. 86 - 94, ISSN Print: 0976 – 6367, ISSN
Online: 0976 – 6375.
30. Apoorv Prem, “Articulated Vehicle Systems” International Journal of Mechanical
Engineering & Technology (IJMET), Volume 5, Issue 7, 2014, pp. 36 - 41, ISSN Print: 0976
– 6340, ISSN Online: 0976 – 6359.

Performance evaluation of different automatic seed point generation techniques for segmentation of indian vehicle number plate

More Related Content

What's hot (16)

Viewers also liked (15)

Similar to Performance evaluation of different automatic seed point generation techniques for segmentation of indian vehicle number plate (20)

More from IAEME Publication (20)

Recently uploaded (20)

Performance evaluation of different automatic seed point generation techniques for segmentation of indian vehicle number plate