Professional Documents
Culture Documents
COMPETITION
COMPETITION
ضوابط برگزاری مسابقات و همچنین شرکت در مسابقات با برگزاری جلسات مختلف با
حضور کمیته مسابقات و دبیر کنفرانس تدوین شده است .این مسابقات به صورت غیر
حضوری و از طریق ثبت نام اولیه و سپس ارسال سیستم های نهایی به ایمیل مسابقات
برگزار شده است .ارزیابی سیستم های ارسال شده با استفاده از نمونه های آزمایشی
در انتهــا ضمــن قدردانــی از حمایــت همــه جانبــه دبیــر کنفرانــس آفــای دکتــر صــدری ﺟﺪول :2ﺗﯿﻢ ﻫﺎي ﺷﺮﮐﺖ ﮐﻨﻨﺪه در ﻣﺴﺎﺑﻘﺎت ﻣﺨﺘﻠﻒ
و تمامــی همــکاران در کمیتــه هــای برگــزاری مســابقات مختلــف کــه تمامــی تــاش رﺗﺒﻪ ﻧﺎم داﻧﺸﮕﺎه ﻧﺎم ﺗﯿﻢ ﻋﻨﻮان ﻣﺴﺎﺑﻘﻪ
خــود را بــرای برگــزاری صحیــح و عادالنــه ایــن مســابقات نمــوده انــد ،امیــدوارم کــه اول داﻧﺸﮕﺎه ﺷﺎﻫﺮود ﺷﺎﻫﺮود OCR
برگــزاری مســابقات در دوره هــای بعــدی کنفرانــس نیــز ادامــه یابــد و تــداوم ایــن
دوم داﻧﺸﮕﺎه ﺑﯿﺮﺟﻨﺪ و ﻫﺮﻣﺰﮔﺎن ﭘﺮدازﺷﮕﺮان ﺑﺎزﺷﻨﺎﺳﯽ ارﻗﺎم
ﺳﻮم داﻧﺸﮕﺎه ﺻﻨﻌﺘﯽ اﻣﯿﺮﮐﺒﯿﺮ اﻣﯿﺮﮐﺒﯿﺮ دﺳﺘﻨﻮﺷﺘﻪ ﻣﺠﺰاي ﻓﺎرﺳﯽ
مســابقات باعــث افزایــش مشــارکت دانشــجویان و محققیــن مختلــف در داخــل و خارج ﭼﻬﺎرم داﻧﺸﮕﺎه ﺷﻬﯿﺪ ﺑﻬﺸﺘﯽ
ﮔﺮوه ﺑﺎزﺷﻨﺎﺳﯽ اﻟﮕﻮي )ﻋﺮﺑﯽ(
کشــور و در نتیجــه پیشــبرد علمــی کشــور عزیزمــان ایــران گــردد. National University of
داﻧﺸﮕﺎه ﺷﻬﯿﺪ ﺑﻬﺸﺘﯽ
اول Singapore
Singapore
دودوﯾﯽ ﻧﻤﻮدن ﺗﺼﺎوﯾﺮ
دوم داﻧﺸﮕﺎه ﺷﻬﯿﺪ ﭼﻤﺮان اﻫﻮاز ﺷﻬﯿﺪ ﭼﻤﺮان
مریم National University of
ﺧﻄﯽ ﻗﺪﯾﻤﯽ ﻓﺎرﺳﯽ
ﺳﻮم Pakistan
خراشادیزاده اول
Sciences and Technology
داﻧﺸﮕﺎه ﻓﺮدوﺳﯽ ﻣﺸﻬﺪ ﭘﺮدازش ﻫﻮﺷﻤﻨﺪ
دبیر کمیته مسابقات اولین کنفرانس بازشناسی الگو و تحلیل تصویر ایران دوم داﻧﺸﮕﺎه ﻫﺮﻣﺰﮔﺎن آرﯾﺎ ﺑﺮرﺳﯽ ﺻﺤﺖ اﻣﻀﺎي
داﻧﺸﮕﺎه آزاد اﺳﻼﻣﯽ -واﺣﺪ ﻋﻠﻮم ﺑﺮون ﺧﻂ ﻓﺎرﺳﯽ
ﺳﻮم ﮐﺎوش
و ﺗﺤﻘﯿﻘﺎت ﺑﯿﺮﺟﻨﺪ
اول داﻧﺸﮕﺎه ﮔﻠﺴﺘﺎن داﻧﺸﮕﺎه ﮔﻠﺴﺘﺎن
دوم داﻧﺸﮕﺎه ﻫﺮﻣﺰﮔﺎن داﻧﺸﮕﺎه ﻫﺮﻣﺰﮔﺎن ﺗﻔﮑﯿﮏ ﺧﻄﻮط در ﻣﺘﻮن
ﺳﻮم داﻧﺸﮕﺎه ﺑﯿﺮﺟﻨﺪ اﻧﺪﯾﺸﻪ ﻧﻮ دﺳﺖ ﻧﻮﯾﺲ
ﭼﻬﺎرم داﻧﺸﮕﺎه ﺻﻨﻌﺘﯽ اﻣﯿﺮﮐﺒﯿﺮ اﻣﯿﺮﮐﺒﯿﺮ
در ایــن مســابقات در مجمــوع 25نفــر در قالــب تیــم هــای مختلــف بــه رقابــت
پرداختنــد کــه برخــی از ایــن شــرکت کننــدگان در بیــش از یــک مســابقه شــرکت
نمودنــد .جــدول 3اطالعــات آمــاری ایــن شــرکت کننــدگان را نشــان مــی دهــد.
همــان طــور کــه در جــدول مشــاهده مــی شــود ،بیشــترین آمــار شــرکت کننــدگان را
دانشــجویان مقطــع کارشناســی ارشــد تشــکیل مــی دهنــد.
ﺟﺪول :3اﻃﻼﻋﺎت ﺷﺮﮐﺖ ﮐﻨﻨﺪﮔﺎن در ﻣﺴﺎﺑﻘﺎت
ﺗﻌﺪاد داﻧﺸﺠﻮﯾﺎن
ﺗﻌﺪاد داﻧﺸﺠﻮﯾﺎن ﺗﻌﺪاد داﻧﺸﺠﻮﯾﺎن ﺗﻌﺪاد اﻋﻀﺎي ﺗﻌﺪاد ﺷﺮﮐﺖ ﺗﻌﺪاد ﮐﻞ
ﻣﻘﻄﻊ ﮐﺎرﺷﻨﺎﺳﯽ
ﻣﻘﻄﻊ ﮐﺎرﺷﻨﺎﺳﯽ ﻣﻘﻄﻊ دﮐﺘﺮي ﻫﯿﺎت ﻋﻠﻤﯽ ﮐﻨﻨﺪﮔﺎن ﺧﺎرﺟﯽ ﺷﺮﮐﺖ ﮐﻨﻨﺪﮔﺎن
ارﺷﺪ
2 12 3 3 5 25
`
q k
Bk
()
(SVD) ICDAR
()
[1] Sadri, J., Izadi, S., Solimanpour, F., Suen, C.Y., Bui, T.D.,
"State-of-the-art in Farsi script recognitions", Int. Symposium on
Signal Processing and its Applications, ISSPA, 2007.
[2] Kosravi, H., Kabir, E.,"Introducing a very large dataset of
Handwritten Farsi digits and a study on their varieties", Pattern
Recognition Letters, Vol. 28, pp. 1133-1141, 2007. OCR
k
KNN
k
OCR
OCR k
k
k
OCR
K
MLP
KNN LBP
s K
k K k
kk
OCR
Gray-scale
LBP
OCR
OCR
MLP
MLP
PRIA 2013
CPU: T7500 @ 2.20 GHZ 2.20 GHZIntel (R) core (TM) 2 Duo
RAM: 3.00 GB
(FLR)
(ECR) []
(RR) (EWR)
[,]
• ICDAR2007 Handwriting Segmentation
Contest []
• ICDAR2009 Handwriting Segmentation
Contest []
• ICFHR 2010 Handwriting Segmentation
Contest []
(Touching)
(slant) (Overlapping)
email: hkm.kardan@gmail.comIT
email:
hhosseini@yazd.ac.ir
IT
email: akbari_younes@yahoo.com
TIF DPI
email: mjavad_jalili@yahoo.comIT
email:mkhorashadi@birjand.ac.irIT
IT
email: jsadri@birjand.ac.ir
(RR)
FLR
ECR
EWR
RR
[1] Alaei, A., Pal, U., Nagabhushan, P.: A new scheme for
unconstrained handwritten text-line segmentation. Patten
Recognition 44(4), 917–928 (2011)
[2] M. Mohammadi, S.S. Mozaffari Chanijani,V.N. Manjunath
Aradhya, G.H. Kumar.: An Improved Handwritten Text Line
Segmentation Technique, pp. 289–296, 2011
[3] A. Zahour, B. Taconet, P. Mercy, S. Ramdane, Arabic hand-
written text-line extraction, in: Proceedings of the Sixth
International Conference on Document Analysis and
Recognition, 2001, pp. 281–285.
[4] A.Zahour, L. Likforman-Sulem, W. Boussalaa, B. Taconet,
Text-line segmenta- tion of historical Arabic documents, in: FLR ECR EWR RR
Proceedings of Ninth International Conference on Document
Analysis and Recognition, 2007, pp. 138–142.
[5] M. Arivazhagan, H. Srinivasan, S.N. Srihari, A statistical
approach to hand- written line segmentation, in: Proceedings of
SPIE Document Recognition and Retrieval XIV, 2007, p.
6500T-1–11.
Abstract—The first competition on the binarization of his- Figure 1 shows sample images and corresponding ground truth
torical Persian documents and manuscripts (PHIBC 2012) has images used in the PHIBC 2012.
been organized in conjunction with the first Iranian conference
on pattern recognition and image analysis (PRIA 2013). The The dataset is available via competition website and tech-
main objective of PHIBC 2012 is to evaluate performance of nical committee 11 website (IAPR TC-11):
the binarization methodologies, when applied on the Persian http://phibc2012.ir/
heritage images. This paper provides a report on the methodology http://www.iapr-tc11.org/
Fig. 1. Sample original and ground truth images used in PHIBC 2012. and performance of the three submitted algorithms based on
evaluation measures has been used. Also, the source code of the evaluation measures used for
performance evaluation can be found at [10].
segmentation based on an average value computed through Keywords—Document image processing, Historical document
1 binarization, Persian heritage manuscripts, Binarization contest The rest of the paper is organized as follows. In section II,
multiple thresholds (based on 4 different Niblack inspired P SN R = 10 × log( ). (2)
M SE a brief description of the submitted methods to PHIBC 2012
thresholding formulas). In final step of post processing, again
is provided. The evaluation measures used for comparison
conditional noise removal and constrained morphological I. I NTRODUCTION
F-Measure can be considered as an intelligent alternative between submitted algorithms are described in section III.
operations are performed to get the final binarised image.
for PSNR because it takes into account the number of fore- Section IV provides experimental results. Finally, section V
There are many old manuscripts and documents in the draws a conclusion.
3-1 Seyed Mehrdad Kankanan and Hossain Poyarad (Faculty ground and background pixels. Let’s T P , F P , F N and T N
libraries and museums of Iran. Many of them include histor-
of engineering, Shahid Chamran University of Ahvaz, Ahvaz, denote the true positive, false positive, false negative and true
ically important data which needs automatic processing and
Iran). negative, respectively. Recall, precision and F-Measure can be II. D ESCRIPTION OF METHODOLOGIES
reading. However, less attention has been made to preserve
defined as:
these valuable types of documents. The PHIBC 2012 is a In the Persian heritage image binarization competition
Proposed method is mainly based on fuzzy measures
primitive attempt toward evaluation of binarization methods, (PHIBC 2012), three groups submitted three algorithms. The
introduced in [11]. It finds a global threshold and apply it TP
Recall = (3) when applied on the Persian manuscripts. description of each methodology is provided by these groups
for whole image. The main advantage of this method against
Otsu’s method, is its better classification of bleed-through in
TP + FN and is as follows.
Persian heritage documents and manuscripts following their
the case of documents that include large amount of this type similarity to Arabic documents [1] are in the form of handwrit-
of degradation. Afterward, an approach similar to Niblack and TP 1- Su Bolan† , Tian Shangxuan† , Lu Shijian‡ and Tan
P recision = (4) ten document images. As usual, for handwritten documents, Chew Lim† († School of Computing, National University of
some post-processing processes are applied to improve the TP + FP preservation of strokes and sub-strokes is of great interest.
final binarization result. Singapore, and ‡ Department of Computer Vision and Image
Previously, five Latin datasets of historical manuscripts has Understanding Institute for Infocomm Research, Singapore).
been publicly available for DIAR researchers [2]–[6]. PHIBC
2 × Recall × P recision
III. E VALUATION MEASURES F − M easure = . (5) 2012 introduces the first dataset that developed for binarization There are four main steps in our proposed method. First,
Recall + P recision of Persian heritage documents. The dataset used for PHIBC local image contrast which is evaluated by local maximum
For an objective comparison between submitted algorithms, 2012 consisted of ten historical documents used with permis- and minimum and local image gradient are combined using an
six evaluation measures used [10]. These are F-Measure [12], Also, pseudo F-Measure is computed like F-Measure ex- sion from ”Documents and old manuscripts treasury of Mirza exponential function with an adaptive factor. Second, The local
pseudo F-Measure [13], peak signal-to-noise ratio (PSNR), dis- cept that recall value is taken from skeletonized ground truth: Mohammad Kazemaini (affiliated with Hazrate Emamzadeh image contrast is combined with the edge map to extract an
tance reciprocal distortion metric (DRD) [14], misclassification Jafar), Yazd, Iran”. The images in PHIBC 2012 suffered accurate text character edge image. Third, the document image
penalty metric (MPM) [15] and negative rate metric (NRM) from various types of degradation, including bleed-through, is binarized by a local threshold which is decided based on
[2], [3]. 2 × Recallskel × P recision
pseudo F − M easure = . (6) deterioration of cellulose structure, faded ink and alien ink, the constructed edge map and estimated stroke width. At last,
Recallskel + P recision among others [7]. For each image in the dataset, a groud some post-processing work is applied to produce better results.
Let’s bwout and GT denotes the binarized image and
ground truth image, respectively. PSNR for binary images can truth is generated with a semi-automatic approach. At first,
Furthermore, NRM can be computed as: 2- Syed Ahsen Raza Ali Hamdani (National University
be defined as: document is processed with phase congruency features used
in [8], [9] to produce a rough binarized image. Then, the final of Sciences and Technology (NUST), Islamabad, Pakistan).
N R F N + N RF P ground truth is generated manually by human expert from the
N M N RM = . (7) The algorithm is based on three processing steps:
2 rough binary image produced in the first step. We will report
x=1 y=1 [GT (x, y) − bwout(x, y)] preprocessing, thresholding and postprocessing. In
M SE = (1) where: the ground truth generation methodology in a dedicated report.
N ×M preprocessing, conditional noise removal and edge based
FN FP processing is performed. Thresholding step involves a
1 Hereinafter, N RF N = , N RF P = (8)
we refer to each group with its assigned number. FN + FP FP + TN c
978-1-4673-6206-1/13/$31.00 2013 IEEE computation of final threshold for background and text
TABLE I. C HARACTERISTICS OF TEST IMAGES USED IN THE PHIBC 2012 DRD measures the distortion for all the S flipped pixels as higher rank. Table II provides detailed experimental results
Image name Size Degradation type(s)
follows: of the binarization algorithms participated in PHIBC 2012
and some state-of-the-art binarization methods. Between three
Persian01 1625×1269 faded ink, multi-degraded background, color background. S participated algorithms, the algorithm submitted by 1- Su
Persian02 845×691 bleed-through, alien ink, low resolution. DRDk
Persian03 1215×735 bleed-through, degraded background, small amount of text. DRD = k=1
. (9) Bolan† , Tian Shangxuan† , Lu Shijian‡ and Tan Chew Lim†
Persian04 1247×1829 faded ink, color background, degraded background. N U BN († School of Computing, National University of Singapore, and
Persian05 887×1149 faded ink, color background, visible fibers in the paper. ‡
Department of Computer Vision and Image Understanding
Persian06 697×1359 faded ink, lines, degraded background. where, DRDk is the distortion of the k − th flipped pixel
Persian07 637×1149 faded ink, lines, degraded background, multi-color background. Institute for Infocomm Research, Singapore) achieved the best
and it is calculated using a 5 × 5 normalized weight matrix
Persian08 1617×969 blur, faded ink, spots, degraded background. performance. Figure 2 shows binarization results of the winner
Persian09 1025×719 faded ink, ink smear, ink noise, alien ink, degraded background. WN m [14]. DRDk equals to the weighted sum of the pixels
Persian10 1649×1258 lines, blur, multi-color background, visible fibers in paper, bleed-through. of PHIBC 2012.
in the 5 × 5 block of the ground truth GT that differ from the
centered k − th flipped pixel at (x, y) in the binarization result
TABLE II. E VALUATION OF THE BINARIZATION METHODS PARTICIPATED IN THE PHIBC 2012 image B.
Method no. Rank / Score F-Measure pseudo F-Measure PSNR DRD MPM(×10−3 ) NRM(×10−2 )
1 1 / 51.3792 88.55 92.25 18.28 5.57 2.33 6.84 2 2
2 3 / 50.2433 86.79 86.29 17.64 6.08 2.74 5.59 DRDk = [GTk (i, j)−Bk (x, y)]×WN m (i, j) (10)
3 2 / 50.7329 87.30 89.50 17.95 5.87 3.79 5.42 i=−2 y=−2
Otsu [16] - 77.75 79.98 15.42 31.11 16.50 5.69
Grid based Sauvola [17] - 85.29 87.75 17.73 9.99 6.01 4.73 NUBN is the number of the non-uniform (not all black
ESBK [18] - 84.03 86.43 17.60 14.79 7.13 5.52 not all white pixels) 8 × 8 blocks in the GT image. MPM is
Su’s method [19] - 88.21 88.82 18.27 5.44 2.65 5.74 a measure of how well the resulting image representing the
PC [8] - 90.19 91.35 19.89 5.23 2.74 3.90 contour of ground truth image and defined as:
Howe [20] - 89.58 91.88 18.53 4.11 2.96 4.56
FN FP
1 i
MPM = ( d + dj ) . (11)
to PHIBC 2012. Based on the performance of the groups ture Notes in Computer Science: Asian Conference on Computer Vision 2D i=1 F N j=1 F P
participated in the competition and state-of-the-art binarization (ACCV’12 Workshops). Springer Berlin / Heidelberg, 2013, vol. 7729,
pp. 1–12.
methodologies, there is a lot of room for development of higher
performance binarization algorithms. [9] H. Ziaei Nafchi and H. Rashidy Kanan, “A phase congruency based where, diF N and djF P denote the distance of the i−th false
document binarization,” in IAPR International Conference on Image positive and the j − th false negative pixel from the contour
and Signal Processing, 2012, pp. 113–121.
ACKNOWLEDGMENT of the text in the ground truth image. The factor D is the sum
[10] R. Farrahi Moghaddam and H. Ziaei Nafchi, “Objective evaluation of
binarization methods,” MATLAB Central File Exchange.
of all the pixel-to-contour distances of the ground truth object.
The organizers of PHIBC 2012 would like to thank ”Doc- http://www.mathworks.com/matlabcentral/fileexchange/27652/, 2013.
uments and old manuscripts treasury of Mirza Mohammad A higher value for F-Measure, pseudo F-Measure and
[11] N. Vieira Lopes, P. A. Mogadouro do Couto, and H. Bustince, “Auto- PSNR measures, indicate to better classification, while a lower
Kazemaini (affiliated with Hazrate Emamzadeh Jafar), Yazd, matic histogram threshold using fuzzy measures,” IEEE Transactions
Iran” for providing us the images used in the PHIBC 2012. on Image Processing, vol. 19, no. 1, pp. 199–204, 2010.
value for DRD, MPM and NRM measures, shows better
[12] M. Sokolova and G. Lapalme, “A systematic analysis of performance
performance.
R EFERENCES measures for classification tasks,” Information Processing and Manage-
ment, vol. 45, pp. 427–437, 2009. IV. E XPERIMENTAL RESULTS
[1] R. Farrahi Moghaddam, M. Cheriet, M. M. Adankan, K. Filonenko, [13] K. Ntirogiannis, B. Gatos, and I. Pratikakis, “An objective evaluation
and R. Wiscovsky, “IBN SINA: A database for research on processing methodology for handwritten image document binarization techniques,” In this section, characteristics of test images used in
and understanding of arabic manuscripts images,” in Document Analysis in Document Analysis Systems, 2008, pp. 217–224. the PHIBC 2012 are described. Images in the PHIBD 2012
Systems, 2010, pp. 11–18. suffered from various types of degradation, include uneven
[14] H. Lu, A. C. Kot, and Y. Q. Shi, “Distance-reciprocal distortion measure
[2] B. Gatos, K. Ntirogiannis, and I. Pratikakis, “ICDAR 2009 document for binary document images,” IEEE Signal Processing Letters, vol. 11, illumination changes, various types of bleed-through, etc. Ta-
image binarization contest (DIBCO 2009),” in International Conference no. 2, pp. 228–231, 2004.
on Document Analysis and Recognition, 2009, pp. 1375–1382. ble I has summarized the degradation types of each document
[15] D. P. Young and J. M. Ferryman, “Pets metrics: On-line performance image used in the PHIBC 2012.
[3] I. Pratikakis, B. Gatos, and K. Ntirogiannis, “H-DIBCO 2010 hand- evaluation service,” in Proceedings of the 14th International Conference
written document image binarization competition,” International Con- on Computer Communications and Networks, 2005, pp. 317–324. For each image in the dataset, best value of each measure
ference on Frontiers in Handwriting Recognition, pp. 727–732, 2010.
[16] N. Otsu, “A threshold selection method from gray-level histograms,” between all of the methods is considered. A method with best
[4] ——, “ICDAR 2011 document image binarization contest (DIBCO IEEE Trans. Systems, Man, and Cybernetics, vol. 9, no. 2, pp. 62–66, Fig. 2. Sample binarization results from the winner of PHIBC 2012.
2011),” in International Conference on Document Analysis and Recog- value for a measure takes a score of 1, and other methods
1979.
nition, 2011, pp. 1506–1510. takes a fraction of 1 by a comparison with the best value.
[17] R. Farrahi Moghaddam and M. Cheriet, “A multi-scale framework for
[5] ——, “ICFHR 2012 competition on handwritten document image adaptive binarization of degraded document images,” Pattern Recogni-
Since there are six evaluation measures and ten test images,
binarization competition,” in International Conference on Frontiers in we can compute the score of a method as: V. C ONCLUSION
tion, vol. 43, no. 6, pp. 2186–2198, 2010.
Handwriting Recognition, 2012, pp. 813–818.
[18] ——, “AdOtsu: An adaptive and parameterless generalization of otsu’s This paper provides a report on the first Persian heritage
[6] R. Rowley-Brook, F. Pitie, and A. Kokaram, “A ground truth bleed- method for document image binarization,” Pattern Recognition, vol. 45,
through document image database,” in International Conference on
10 6
Besti,j valuek,i,j image binarization competition (PHIBC 2012) which has been
pp. 2419–2431, 2012. 3
(12)
Theory and Practice of Digital Libraries, 2012, pp. 185–196. Sk=1 = ( , ). organized in conjunction with the first Iranian conference on
[19] B. Su, S. Lu, and C. Tan, “Binarization of historical document im- valuek,i,j Besti,j
[7] “Special issue on recent advances in applications to visual cultural i=1 j=1 pattern recognition and image analysis (PRIA 2013). The main
ages using the local maximum and minimum,” in Document Analysis
heritage,” IEEE Signal Processing Magazine, vol. 12, no. 1, pp. 234– Systems, 2010, pp. 159–166. objective of this competition is to evaluate the performance
778, 2008.
[20] N. Howe, “A laplacian energy for document binarization,” in Interna- where, k denotes the number of participators and value of the binarization methods, when applied on the historical
[8] H. Ziaei Nafchi, R. Farrahi Moghaddam, and M. Cheriet, “Historical tional Conference on Document Analysis and Recognition, 2011, pp. is the measure value obtained by a method. First fraction is Persian document images. The images used in PHIBC 2012
document binarization based on phase information of images,” in Lec- 6–10. used for those measures in which a lower value indicates to include wide range of degradation types and their associated
better score, and the second one is used for measures with ground truth are publicly available. Six evaluation measures
inverse behavior. Finally, methods with higher scores take has been used for comparison between submitted algorithms
PRIA 2013
(FRR)(FAR)
[,,,]
off-line
FRRFAR
(SVC 2004) •
(FAR) [] FAR: False Acceptance Rate
(FRR)
FRR: False Rejection Rate
• PDA: Personal Digital Assistance
(AER) [](ICDAR 2009) AER: Average Error Rate
•
[](BSEC 2009)
•
(ICFHR 2010) [,]
[]
.
•
[](ICDAR 2011)
(PDA)
•
(ICFHR 2012)
[]
email: akbari_younes@yahoo.com IT
email: mjavad_jalili@yahoo.comIT
email: km_farda2006@yahoo.com IT
email:mkhorashadi@birjand.ac.irIT
IT
email:jsadri@birjand.ac.ir
Workshop on Automated Forensic Handwriting Analysis
IWAFHA (2011) pp. 15.
[7] S. Ghandali and M. E. Moghaddam, “Off-line Persian signature
identification and verification cation based on image registration FAR(%) FRR(%) AER(%)
and fusion,” J. Multimed., 4(3) (2009) 137-144.
[8] M. Radmehr, S. M. Anisheh, M. Nikpour and A. Yaseri,
“Designing an off-line method for signature recognition,”
World Appl. Sci. J. 13(3) (2011) 438-443.
[9] Atefeh Foroozandeh, Younes Akbari, Mohammad Javad Jalili,
Javad Sadri, “A Novel and Practical System for Verifying
Signatures on Persian Handwritten Bank Checks”, Accepted for
publication in the International Journal of Pattern Recognition
and Artificial Intelligence (IJPRAI),
Vol.26,No.6(2012)(27pages),doi:
10.1142/S0218001412560149.
[10] Atefeh Foroozandeh, Mohammad J. Jalili, Younes Akbari,
and Javad Sadri, “Persian Signature Verification Based on
Fractal Dimension Using Testing Hypothesis", Published in
the Proceedings of the 2012 International Conference on
Frontiers in Handwriting Recognition (ICFHR-2012, Italy) on
18-20 Sept. 2012, at Porto Giardino, Bari, Italy.
MLP
[11] D. yan Yeung, H. Chang, Y. Xiong, S. George, R. Kashi,
T. Matsumoto, and G. Rigoll, “SVC2004: First international
signature verification competition,” in In Int. Conf. on Bio-
metric Authentication (ICBA), Hong Kong. Springer, 2004,
pp. 16–22.
[12] N. Houmani, A. Mayoue, and et al., “BioSecure signature
evaluation campaign (BSEC’2009): Evaluating online
signaturealgorithms depending on the quality of signatures,”
Pattern Recognition, Aug. 2011.
[13] V. L. Blankers, C. E. van den Heuvel, K. Y. Franke, and L. G.
Vuurpijl, “ICDAR 2009 signature verification competition,”
inICDAR, 2009, pp. 1403–1407.
[14] M. Liwicki, C. E. van den Heuvel, B. Found, and M. I. Malik, [1] R. Plamondon and G. Lorette, “Automatic signature verification
“Forensic signature verification competition 4NSigComp2010 and writer identification – the state of the art,” in Pattern
- detection of simulated and disguised signatures,” in ICFHR, Recognition, vol. 22, 1989, pp. 107–131.
November 16-18, India, 2010, pp. 715–720. [2] F. Leclerc and R. Plamondon, “Automatic signature
verification: the state of the art 1989–1993,” in Progress in
[15] M. Liwicki, M. I. Malik, C. E. van den Heuvel, X. Chen,C. Automatic Signature Verification, R. Plamondon, Ed. World (FRR) (FAR)
Berger, R. Stoel, M. Blumenstein, and B. Found, “Signature Scientific Publ. Co., 1994, pp. 13–19.
verification competition for online and offline skilled [3] V. Nguyen, M. Blumenstein and G. Leedham, “Global
forgeries SigComp2011,” in ICDAR, 2011, pp. 1480–1484. features for the off-line signature verification problem, ” in
Proc. 10th Int. Conf. on Document Analysis and Recognition
[16] M. Liwicki, M. I. Malik, L. Alewijnse, E. Heuvel and ICDAR (2009) 1300-1304. FAR=
B. found“ICFHR2012 Competition onAutomatic
Forensic Signature Verification (4NsigComp 2012),” [4] R. Plamondon and S. N. Srihari, “On-line and off-line
handwriting recognition: a comprehensive survey,” IEEE FRR=
Published in the Proceedings of the 2012 International
Conference on Frontiers in Handwriting Recognition (ICFHR- Trans.Pattern Analysis and Machine Intelligence, vol. 22, no. 1,
2012, Italy) on 18-20 Sept. 2012, at Porto Giardino, Bari, pp. 63–84, 2000.
Italy.
[5] D. Impedovo and G. Pirlo, “Automatic signature verification:
The state of the art,” IEEE Transactions on Systems, Man,
and Cybernetics, Part C (Applications and Reviews), vol. 38,
no. 5, pp. 609–635, Sep. 2008.