International Journal of Science and Research (IJSR), 2015
This paper discusses the different methods for optical character recognition (OCR), which has bee... more This paper discusses the different methods for optical character recognition (OCR), which has been an important field to research from a few decades due its huge necessity to convert paper documents or books in computer readable format. Though Bangla (widely used as Bengali) is one of the top uses language among the other languages, but there is no reliable character recognizer for this. Our work has covered a total process to develop a complete OCR, especially for feature extraction process, which is very important to recognize characters correctly. Here, we have developed and tested many algorithms to identify each ones merits and limitations in various cases for hand written character recognition to make the stage more optimized. Moreover, we have used hidden Markov model (HMM) classifier along with artificial neural network (ANN) to make our classifier more accurate.
2012 10th IAPR International Workshop on Document Analysis Systems, 2012
Text detection in video frames plays a vital role in enhancing the performance of information ext... more Text detection in video frames plays a vital role in enhancing the performance of information extraction systems because the text in video frames helps in indexing and retrieving video efficiently and accurately. This paper presents a new method for arbitrarily-oriented text detection in video, based on dominant text pixel selection, text representatives and region growing. The method uses gradient pixel direction and magnitude corresponding to Sobel edge pixels of the input frame to obtain dominant text pixels. Edge components in the Sobel edge map corresponding to dominant text pixels are then extracted and we call them text representatives. We eliminate broken segments of each text representatives to get candidate text representatives. Then the perimeter of candidate text representatives grows along the text direction in the Sobel edge map to group the neighboring text components which we call word patches. The word patches are used for finding the direction of text lines and then the word patches are expanded in the same direction in the Sobel edge map to group the neighboring word patches and to restore missing text information. This results in extraction of arbitrarilyoriented text from the video frame. To evaluate the method, we considered arbitrarily-oriented data, non-horizontal data, horizontal data, Hua's data and ICDAR-2003 competition data (Camera images). The experimental results show that the proposed method outperforms the existing method in terms of recall and f-measure.
2012 10th IAPR International Workshop on Document Analysis Systems, 2012
Musical staff line detection and removal techniques detect the staff positions in musical documen... more Musical staff line detection and removal techniques detect the staff positions in musical documents and segment musical score from musical documents by removing those staff lines. It is an important preprocessing step for ensuing the Optical Music Recognition tasks. This paper proposes an effective staff line detection and removal method that makes use of the global information of the musical document and models the staff line shape. It first estimates the staff height and space, and then models the shape of the staff line by examining the orientation of the staff pixels. At last the estimated model is used to find out the location of staff lines and hence to remove those detected staff lines. The proposed technique is simple, robust, and involves few parameters. It has been tested on the dataset of the recent staff removal competition [1] held under the International Conference of Document Analysis and Recognition(ICDAR) 2011. Experimental results show the effectiveness and robustness of our proposed technique on musical documents with various types of deformations.
Advances in Intelligent Systems and Computing, 2013
Biometric systems play a significant role in the field of information security as they are extrem... more Biometric systems play a significant role in the field of information security as they are extremely required for user authentication. Signature identification and verification have a great importance for authentication intention. The purpose of this paper is to present an empirical contribution towards the understanding of multi-script (Hindi and English) signature verification. This system will identify whether a claimed signature belongs to the group of English signatures or Hindi signatures from a combined Hindi and English signature datasets and then it will verify signatures using these two resultant signature datasets (Hindi script signature and English script signatures) separately. The modified gradient feature and SVM classifier were employed for identification and verification purposes. To the best of authors' knowledge, the multi-script signature identification and verification has never been used for the task of signature verification and this is the first report of using Hindi and English signatures in this area. Two different results for identification and verification are calculated and analysed. The accuracy of 98.05% is obtained for the identification of signature script using 2160 (1080 Hindi + 1080 English) samples for training and 1080 (540 Hindi + 540 English) samples for testing. The resultant data sets obtained in script identification of signatures were used for verification purpose. The FRR, FAR for Hindi and English was obtained 8.0%, 4.0% and 12.0%, 10.0% respectively.
Seventh International Conference on Document Analysis and Recognition, 2003. Proceedings.
To take care of variability involved in the writing style of different individuals in this paper ... more To take care of variability involved in the writing style of different individuals in this paper we propose a robust scheme to segment unconstrained handwritten Bangla texts into lines, words and characters. For line segmentation, at first, we divide the text into vertical stripes. Stripe width of a document is computed by statistical analysis of the text height in the document. Next we determine horizontal histogram of these stripes and the relationship of the minimal values of the histograms is used to segment text lines. Based on vertical projection profile lines are segmented into words. Segmentation of characters from handwritten word is very tricky as the characters are seldom vertically separable. We use a concept based on water reservoir principle for the purpose. Here we, at first, identify isolated and connected (touching) characters in a word. Next touching characters of the word are segmented based on the reservoir base area points and structural feature of the component.
2013 2nd IAPR Asian Conference on Pattern Recognition, 2013
This paper presents a survey on sclera-based biometric recognition. Among the various biometric m... more This paper presents a survey on sclera-based biometric recognition. Among the various biometric methods, sclera is one of the novel and promising biometric techniques. The sclera, a white region of connective tissue and blood vessels, surrounds the iris. A survey of the techniques available in the area of sclera biometrics will be of great assistance to researchers, and hence a comprehensive effort is made in this article to discuss the advancements reported in this regard during the past few decades. As a limited number of publications are found in the literature, an attempt is made in this paper to increase awareness of this area so that the topic gains popularity and interest among researchers. In this survey, a brief introduction is given initially about the sclera biometric, which is subsequently followed by background concepts, various pre-processing techniques, feature extraction and finally classification techniques associated with the sclera biometric. Benchmarking databases are very important for any pattern recognition related research, so the databases related with this work is also discussed. Finally, our observations, future scope and existing difficulties, which are unsolved in sclera biometrics, are discussed. We hope that this survey will serve to focus more researcher attention towards the emerging sclera biometric.
Automatic separation of text and symbols from graphics in document image is one of the fundamenta... more Automatic separation of text and symbols from graphics in document image is one of the fundamental aims in graphics recognition. In maps, separation of text and symbols from graphics involves many challenges because the text and symbols frequently touch/overlap with graphical components. Sometimes the colors in a single character are gradually distributed which adds extra difficulty in text and symbol separation from color maps. In this paper we proposed a system to retrieve text and symbol from color map. Here, at first, we separate the map into different foreground layers according to color features and then in each layer, connected component features and skeleton information are used to identify text and symbol from graphics on the basis of their geometrical features. Lastly, segmentation results of the individual layers are combined to get final segmentation results. From the experiment we obtained encouraging results.
2014 14th International Conference on Frontiers in Handwriting Recognition, 2014
1,000 music pages written by 50 different writers. Every writer has 20 different music pages.... more 1,000 music pages written by 50 different writers. Every writer has 20 different music pages. Dataset divided in two parts for training and testing.
International Journal of Innovative Technology and Exploring Engineering, 2019
Segmentation is division of something into smaller parts and one of the Component of character re... more Segmentation is division of something into smaller parts and one of the Component of character recognition system. Separation of characters, words and lines are done in Segmentation from text documents. character recognition is a process which allows computers to recognize written or printed characters such as numbers or letters and to change them into a form that the computer can use. the accuracy of OCR system is done by taking the output of an OCR run for an image and comparing it to the original version of the same text. The main aim of this paper is to find out the various text line segmentations are Projection profiles, Weighted Bucket Method. Proposed method is horizontal projection profile and connected component method on Handwritten Kannada language. These methods are used for experimentation and finally comparing their accuracy and results.
2013 12th International Conference on Document Analysis and Recognition, 2013
This paper presents a two-stage method for multioriented video character segmentation. Words segm... more This paper presents a two-stage method for multioriented video character segmentation. Words segmented from video text lines are considered for character segmentation in the present work. Words can contain isolated or non-touching characters, as well as touching characters. Therefore, the character segmentation problem can be viewed as a two stage problem. In the first stage, text cluster is identified and isolated (nontouching) characters are segmented. The orientation of each word is computed and the segmentation paths are found in the direction perpendicular to the orientation. Candidate segmentation points computed using the top distance profile are used to find the segmentation path between the characters considering the background cluster. In the second stage, the segmentation results are verified and a check is performed to ascertain whether the word component contains touching characters or not. The average width of the components is used to find the touching character components. For segmentation of the touching characters, segmentation points are then found using average stroke width information, along with the top and bottom distance profiles. The proposed method was tested on a large dataset and was evaluated in terms of precision, recall and f-measure. A comparative study with existing methods reveals the superiority of the proposed method.
2012 International Conference on Digital Image Computing Techniques and Applications (DICTA), 2012
Word segmentation has become a research topic to improve OCR accuracy for video text recognition,... more Word segmentation has become a research topic to improve OCR accuracy for video text recognition, because a video text line suffers from arbitrary orientation, complex background and low resolution. Therefore, for word segmentation from arbitrarily-oriented video text lines, in this paper, we extract four new gradient directional features for each Canny edge pixel of the input text line image to produce four respective pixel candidate images. The union of four pixel candidate images is performed to obtain a text candidate image. The sequence of the components in the text candidate image according to the text line is determined using nearest neighbor criteria. Then we propose a two-stage method for segmenting words. In the first stage, for the distances between the components, we apply K-means clustering with K=2 to get probable word and non-word spacing clusters. The words are segmented based on probable word spacing and all other components are passed to the second stage for segmenting correct words. For each segmented and un-segmented words passed to the second stage, the method repeats all the steps until the K-means clustering step to find probable word and non-word spacing clusters. Then the method considers cluster nature, height and width of the components to identify the correct word spacing. The method is tested extensively on video curved text lines, non-horizontal straight lines, horizontal straight lines and text lines from the ICDAR-2003 competition data. Experimental results and a comparative study shows the results are encouraging and promising.
Proceedings of Sixth International Conference on Document Analysis and Recognition
In U ggnerul situation, U document page may contuin severul script forms. For Optical Churucter R... more In U ggnerul situation, U document page may contuin severul script forms. For Optical Churucter Recognition (OCR) of such U document page, it is necessuiy to sepurute the scripts before feeding them to their individuul OCR systems. In this puper, un uutomutic technique for the ...
Proceedings of 2005 International Conference on Intelligent Sensing and Information Processing, 2005.
... there is no significallt work This paper deals with word-wise handwritten script identificati... more ... there is no significallt work This paper deals with word-wise handwritten script identification for Indian postal automation. In the proposed scheme at first document skew is detected and corrected. NOli-text parts are then segmented finm the document USing RIm Length ...
There are many artistic documents where text lines of a single page may have different inclinatio... more There are many artistic documents where text lines of a single page may have different inclinations (orientations). To enhance the ability of document analysis system, we have to extract text line in multiple orientations. In this paper, we propose a robust technique to detect English text lines of arbitrary orientation in a single document page. We propose here a bottom-up approach where the connected components are at first labelled. They are then clustered into word groups. Text lines of arbitrary orientation are identified from the estimation of these word groups. From an experiment of 3700 text lines, we obtained an accuracy of 98.3% by the proposed method.
International Journal on Document Analysis and Recognition (IJDAR), 2015
Thinning that preserves visual topology of characters in video is challenging in the field of doc... more Thinning that preserves visual topology of characters in video is challenging in the field of document analysis and video text analysis due to low resolution and complex background. This paper proposes to explore ring radius transform (RRT) to generate a radius map from Canny edges of each input image to obtain its medial axis. A radius value contained in the radius map here is the nearest distance to the edge pixels on contours. For the radius map, the method proposes a novel idea for identifying medial axis (middle pixels between two strokes) for arbitrary orientations of the character. Iterative-maximal-growing is then proposed to connect missing medial axis pixels at junctions and intersections. Next, we perform histogram on color information of medial axes with clustering to eliminate false medial axis segments. The method finally restores the shape of the character through radius values of medial axis pixels for the purpose of recognition with the Google Open source OCR (Tesseract). The method has been tested on video, natural scene and handwrit
The purpose of this paper is to present an empirical contribution towards the understanding of mu... more The purpose of this paper is to present an empirical contribution towards the understanding of multi-script off-line signature identification and verification using a novel method involving off-line Hindi (Devnagari) and English signatures. The main aim of this approach is to demonstrate the significant advantage of the use of signature script identification in a multiscript signature verification environment. In the 1 st stage of the proposed signature verification technique a script identification technique is employed to know whether a signature is written in Hindi or English. In the second stage, a verification approach was explored separately for English signatures and Hindi signatures based on the script identification result. Different features like gradient feature, water reservoir feature, loop feature, aspect ratio etc. were employed, and Support Vector Machines (SVMs) were considered in our scheme. To get the comparative idea, multi-script signature verification results on the joint Hindi and English dataset without using any script identification technique is also computed. From the experiment results it is noted that we are able to reduce average error rate 4.81% more when script identification method is employed.
The 2012 International Joint Conference on Neural Networks (IJCNN), 2012
In the field of information security, the usage of biometrics is growing for user authentication.... more In the field of information security, the usage of biometrics is growing for user authentication. Automatic signature recognition and verification is one of the biometric techniques, which is only one of several used to verify the identity of individuals. In this paper, a foreground and background based technique is proposed for identification of scripts from bilingual (English/Roman and Chinese) off-line signatures. This system will identify whether a claimed signature belongs to the group of English signatures or Chinese signatures. The identification of signatures based on its script is a major contribution for multi-script signature verification. Two background information extraction techniques are used to produce the background components of the signature images. Gradient-based method was used to extract the features of the foreground as well as background components. Zernike Moment feature was also employed on signature samples. Support Vector Machine (SVM) is used as the classifier for signature identification in the proposed system. A database of 1120 (640 English+480 Chinese) signature samples were used for training and 560 (320 English+240 Chinese) signature samples were used for testing the proposed system. An encouraging identification accuracy of 97.70% was obtained using gradient feature from the experiment.
2009 10th International Conference on Document Analysis and Recognition, 2009
In recent years research towards Indian handwritten character recognition is getting increasing a... more In recent years research towards Indian handwritten character recognition is getting increasing attention. Many approaches have been proposed by the researchers towards handwritten Indian character recognition and many recognition systems for isolated handwritten numerals/characters are available in the literature. To get idea of the recognition results of different classifiers and to provide new benchmark for future research, in this paper a comparative study of Devnagari handwritten character recognition using twelve different classifiers and four sets of feature is presented. Projection distance, subspace method, linear discriminant function, support vector machines, modified quadratic discriminant function, mirror image learning, Euclidean distance, nearest neighbour, k-Nearest neighbour, modified projection distance, compound projection distance, and compound modified quadratic discriminant function are used as different classifiers. Feature sets used in the classifiers are computed based on curvature and gradient information obtained from binary as well as gray-scale images.
2014 22nd International Conference on Pattern Recognition, 2014
This work addresses the problem of creating a Bayesian Network based online semi-supervised handw... more This work addresses the problem of creating a Bayesian Network based online semi-supervised handwritten character recognisor, which learns continuously over time to make a adaptable recognisor. The proposed method makes learning possible from a continuous inflow of a potentially unlimited amount of data without the requirement for storage. It highlights the use of unlabelled data for boosting the accuracy, especially when labelled data is scarce and expensive unlike unlabelled data. An algorithm is introduced to perform semi-supervised learning based on the combination of novel online ensemble of the Randomized Bayesian network classifiers and a novel online variant of the Expectation Maximization (EM) algorithm. We make use of a novel varying weighting factor to modulate the contribution of unlabelled data. Proposed method was evaluated using online handwritten Tamil characters from the IWFHR 2006 competition dataset. The accuracy obtained was comparable to the state of the art batch learning methods like HMM and SVMs.
International Journal of Science and Research (IJSR), 2015
This paper discusses the different methods for optical character recognition (OCR), which has bee... more This paper discusses the different methods for optical character recognition (OCR), which has been an important field to research from a few decades due its huge necessity to convert paper documents or books in computer readable format. Though Bangla (widely used as Bengali) is one of the top uses language among the other languages, but there is no reliable character recognizer for this. Our work has covered a total process to develop a complete OCR, especially for feature extraction process, which is very important to recognize characters correctly. Here, we have developed and tested many algorithms to identify each ones merits and limitations in various cases for hand written character recognition to make the stage more optimized. Moreover, we have used hidden Markov model (HMM) classifier along with artificial neural network (ANN) to make our classifier more accurate.
2012 10th IAPR International Workshop on Document Analysis Systems, 2012
Text detection in video frames plays a vital role in enhancing the performance of information ext... more Text detection in video frames plays a vital role in enhancing the performance of information extraction systems because the text in video frames helps in indexing and retrieving video efficiently and accurately. This paper presents a new method for arbitrarily-oriented text detection in video, based on dominant text pixel selection, text representatives and region growing. The method uses gradient pixel direction and magnitude corresponding to Sobel edge pixels of the input frame to obtain dominant text pixels. Edge components in the Sobel edge map corresponding to dominant text pixels are then extracted and we call them text representatives. We eliminate broken segments of each text representatives to get candidate text representatives. Then the perimeter of candidate text representatives grows along the text direction in the Sobel edge map to group the neighboring text components which we call word patches. The word patches are used for finding the direction of text lines and then the word patches are expanded in the same direction in the Sobel edge map to group the neighboring word patches and to restore missing text information. This results in extraction of arbitrarilyoriented text from the video frame. To evaluate the method, we considered arbitrarily-oriented data, non-horizontal data, horizontal data, Hua's data and ICDAR-2003 competition data (Camera images). The experimental results show that the proposed method outperforms the existing method in terms of recall and f-measure.
2012 10th IAPR International Workshop on Document Analysis Systems, 2012
Musical staff line detection and removal techniques detect the staff positions in musical documen... more Musical staff line detection and removal techniques detect the staff positions in musical documents and segment musical score from musical documents by removing those staff lines. It is an important preprocessing step for ensuing the Optical Music Recognition tasks. This paper proposes an effective staff line detection and removal method that makes use of the global information of the musical document and models the staff line shape. It first estimates the staff height and space, and then models the shape of the staff line by examining the orientation of the staff pixels. At last the estimated model is used to find out the location of staff lines and hence to remove those detected staff lines. The proposed technique is simple, robust, and involves few parameters. It has been tested on the dataset of the recent staff removal competition [1] held under the International Conference of Document Analysis and Recognition(ICDAR) 2011. Experimental results show the effectiveness and robustness of our proposed technique on musical documents with various types of deformations.
Advances in Intelligent Systems and Computing, 2013
Biometric systems play a significant role in the field of information security as they are extrem... more Biometric systems play a significant role in the field of information security as they are extremely required for user authentication. Signature identification and verification have a great importance for authentication intention. The purpose of this paper is to present an empirical contribution towards the understanding of multi-script (Hindi and English) signature verification. This system will identify whether a claimed signature belongs to the group of English signatures or Hindi signatures from a combined Hindi and English signature datasets and then it will verify signatures using these two resultant signature datasets (Hindi script signature and English script signatures) separately. The modified gradient feature and SVM classifier were employed for identification and verification purposes. To the best of authors' knowledge, the multi-script signature identification and verification has never been used for the task of signature verification and this is the first report of using Hindi and English signatures in this area. Two different results for identification and verification are calculated and analysed. The accuracy of 98.05% is obtained for the identification of signature script using 2160 (1080 Hindi + 1080 English) samples for training and 1080 (540 Hindi + 540 English) samples for testing. The resultant data sets obtained in script identification of signatures were used for verification purpose. The FRR, FAR for Hindi and English was obtained 8.0%, 4.0% and 12.0%, 10.0% respectively.
Seventh International Conference on Document Analysis and Recognition, 2003. Proceedings.
To take care of variability involved in the writing style of different individuals in this paper ... more To take care of variability involved in the writing style of different individuals in this paper we propose a robust scheme to segment unconstrained handwritten Bangla texts into lines, words and characters. For line segmentation, at first, we divide the text into vertical stripes. Stripe width of a document is computed by statistical analysis of the text height in the document. Next we determine horizontal histogram of these stripes and the relationship of the minimal values of the histograms is used to segment text lines. Based on vertical projection profile lines are segmented into words. Segmentation of characters from handwritten word is very tricky as the characters are seldom vertically separable. We use a concept based on water reservoir principle for the purpose. Here we, at first, identify isolated and connected (touching) characters in a word. Next touching characters of the word are segmented based on the reservoir base area points and structural feature of the component.
2013 2nd IAPR Asian Conference on Pattern Recognition, 2013
This paper presents a survey on sclera-based biometric recognition. Among the various biometric m... more This paper presents a survey on sclera-based biometric recognition. Among the various biometric methods, sclera is one of the novel and promising biometric techniques. The sclera, a white region of connective tissue and blood vessels, surrounds the iris. A survey of the techniques available in the area of sclera biometrics will be of great assistance to researchers, and hence a comprehensive effort is made in this article to discuss the advancements reported in this regard during the past few decades. As a limited number of publications are found in the literature, an attempt is made in this paper to increase awareness of this area so that the topic gains popularity and interest among researchers. In this survey, a brief introduction is given initially about the sclera biometric, which is subsequently followed by background concepts, various pre-processing techniques, feature extraction and finally classification techniques associated with the sclera biometric. Benchmarking databases are very important for any pattern recognition related research, so the databases related with this work is also discussed. Finally, our observations, future scope and existing difficulties, which are unsolved in sclera biometrics, are discussed. We hope that this survey will serve to focus more researcher attention towards the emerging sclera biometric.
Automatic separation of text and symbols from graphics in document image is one of the fundamenta... more Automatic separation of text and symbols from graphics in document image is one of the fundamental aims in graphics recognition. In maps, separation of text and symbols from graphics involves many challenges because the text and symbols frequently touch/overlap with graphical components. Sometimes the colors in a single character are gradually distributed which adds extra difficulty in text and symbol separation from color maps. In this paper we proposed a system to retrieve text and symbol from color map. Here, at first, we separate the map into different foreground layers according to color features and then in each layer, connected component features and skeleton information are used to identify text and symbol from graphics on the basis of their geometrical features. Lastly, segmentation results of the individual layers are combined to get final segmentation results. From the experiment we obtained encouraging results.
2014 14th International Conference on Frontiers in Handwriting Recognition, 2014
1,000 music pages written by 50 different writers. Every writer has 20 different music pages.... more 1,000 music pages written by 50 different writers. Every writer has 20 different music pages. Dataset divided in two parts for training and testing.
International Journal of Innovative Technology and Exploring Engineering, 2019
Segmentation is division of something into smaller parts and one of the Component of character re... more Segmentation is division of something into smaller parts and one of the Component of character recognition system. Separation of characters, words and lines are done in Segmentation from text documents. character recognition is a process which allows computers to recognize written or printed characters such as numbers or letters and to change them into a form that the computer can use. the accuracy of OCR system is done by taking the output of an OCR run for an image and comparing it to the original version of the same text. The main aim of this paper is to find out the various text line segmentations are Projection profiles, Weighted Bucket Method. Proposed method is horizontal projection profile and connected component method on Handwritten Kannada language. These methods are used for experimentation and finally comparing their accuracy and results.
2013 12th International Conference on Document Analysis and Recognition, 2013
This paper presents a two-stage method for multioriented video character segmentation. Words segm... more This paper presents a two-stage method for multioriented video character segmentation. Words segmented from video text lines are considered for character segmentation in the present work. Words can contain isolated or non-touching characters, as well as touching characters. Therefore, the character segmentation problem can be viewed as a two stage problem. In the first stage, text cluster is identified and isolated (nontouching) characters are segmented. The orientation of each word is computed and the segmentation paths are found in the direction perpendicular to the orientation. Candidate segmentation points computed using the top distance profile are used to find the segmentation path between the characters considering the background cluster. In the second stage, the segmentation results are verified and a check is performed to ascertain whether the word component contains touching characters or not. The average width of the components is used to find the touching character components. For segmentation of the touching characters, segmentation points are then found using average stroke width information, along with the top and bottom distance profiles. The proposed method was tested on a large dataset and was evaluated in terms of precision, recall and f-measure. A comparative study with existing methods reveals the superiority of the proposed method.
2012 International Conference on Digital Image Computing Techniques and Applications (DICTA), 2012
Word segmentation has become a research topic to improve OCR accuracy for video text recognition,... more Word segmentation has become a research topic to improve OCR accuracy for video text recognition, because a video text line suffers from arbitrary orientation, complex background and low resolution. Therefore, for word segmentation from arbitrarily-oriented video text lines, in this paper, we extract four new gradient directional features for each Canny edge pixel of the input text line image to produce four respective pixel candidate images. The union of four pixel candidate images is performed to obtain a text candidate image. The sequence of the components in the text candidate image according to the text line is determined using nearest neighbor criteria. Then we propose a two-stage method for segmenting words. In the first stage, for the distances between the components, we apply K-means clustering with K=2 to get probable word and non-word spacing clusters. The words are segmented based on probable word spacing and all other components are passed to the second stage for segmenting correct words. For each segmented and un-segmented words passed to the second stage, the method repeats all the steps until the K-means clustering step to find probable word and non-word spacing clusters. Then the method considers cluster nature, height and width of the components to identify the correct word spacing. The method is tested extensively on video curved text lines, non-horizontal straight lines, horizontal straight lines and text lines from the ICDAR-2003 competition data. Experimental results and a comparative study shows the results are encouraging and promising.
Proceedings of Sixth International Conference on Document Analysis and Recognition
In U ggnerul situation, U document page may contuin severul script forms. For Optical Churucter R... more In U ggnerul situation, U document page may contuin severul script forms. For Optical Churucter Recognition (OCR) of such U document page, it is necessuiy to sepurute the scripts before feeding them to their individuul OCR systems. In this puper, un uutomutic technique for the ...
Proceedings of 2005 International Conference on Intelligent Sensing and Information Processing, 2005.
... there is no significallt work This paper deals with word-wise handwritten script identificati... more ... there is no significallt work This paper deals with word-wise handwritten script identification for Indian postal automation. In the proposed scheme at first document skew is detected and corrected. NOli-text parts are then segmented finm the document USing RIm Length ...
There are many artistic documents where text lines of a single page may have different inclinatio... more There are many artistic documents where text lines of a single page may have different inclinations (orientations). To enhance the ability of document analysis system, we have to extract text line in multiple orientations. In this paper, we propose a robust technique to detect English text lines of arbitrary orientation in a single document page. We propose here a bottom-up approach where the connected components are at first labelled. They are then clustered into word groups. Text lines of arbitrary orientation are identified from the estimation of these word groups. From an experiment of 3700 text lines, we obtained an accuracy of 98.3% by the proposed method.
International Journal on Document Analysis and Recognition (IJDAR), 2015
Thinning that preserves visual topology of characters in video is challenging in the field of doc... more Thinning that preserves visual topology of characters in video is challenging in the field of document analysis and video text analysis due to low resolution and complex background. This paper proposes to explore ring radius transform (RRT) to generate a radius map from Canny edges of each input image to obtain its medial axis. A radius value contained in the radius map here is the nearest distance to the edge pixels on contours. For the radius map, the method proposes a novel idea for identifying medial axis (middle pixels between two strokes) for arbitrary orientations of the character. Iterative-maximal-growing is then proposed to connect missing medial axis pixels at junctions and intersections. Next, we perform histogram on color information of medial axes with clustering to eliminate false medial axis segments. The method finally restores the shape of the character through radius values of medial axis pixels for the purpose of recognition with the Google Open source OCR (Tesseract). The method has been tested on video, natural scene and handwrit
The purpose of this paper is to present an empirical contribution towards the understanding of mu... more The purpose of this paper is to present an empirical contribution towards the understanding of multi-script off-line signature identification and verification using a novel method involving off-line Hindi (Devnagari) and English signatures. The main aim of this approach is to demonstrate the significant advantage of the use of signature script identification in a multiscript signature verification environment. In the 1 st stage of the proposed signature verification technique a script identification technique is employed to know whether a signature is written in Hindi or English. In the second stage, a verification approach was explored separately for English signatures and Hindi signatures based on the script identification result. Different features like gradient feature, water reservoir feature, loop feature, aspect ratio etc. were employed, and Support Vector Machines (SVMs) were considered in our scheme. To get the comparative idea, multi-script signature verification results on the joint Hindi and English dataset without using any script identification technique is also computed. From the experiment results it is noted that we are able to reduce average error rate 4.81% more when script identification method is employed.
The 2012 International Joint Conference on Neural Networks (IJCNN), 2012
In the field of information security, the usage of biometrics is growing for user authentication.... more In the field of information security, the usage of biometrics is growing for user authentication. Automatic signature recognition and verification is one of the biometric techniques, which is only one of several used to verify the identity of individuals. In this paper, a foreground and background based technique is proposed for identification of scripts from bilingual (English/Roman and Chinese) off-line signatures. This system will identify whether a claimed signature belongs to the group of English signatures or Chinese signatures. The identification of signatures based on its script is a major contribution for multi-script signature verification. Two background information extraction techniques are used to produce the background components of the signature images. Gradient-based method was used to extract the features of the foreground as well as background components. Zernike Moment feature was also employed on signature samples. Support Vector Machine (SVM) is used as the classifier for signature identification in the proposed system. A database of 1120 (640 English+480 Chinese) signature samples were used for training and 560 (320 English+240 Chinese) signature samples were used for testing the proposed system. An encouraging identification accuracy of 97.70% was obtained using gradient feature from the experiment.
2009 10th International Conference on Document Analysis and Recognition, 2009
In recent years research towards Indian handwritten character recognition is getting increasing a... more In recent years research towards Indian handwritten character recognition is getting increasing attention. Many approaches have been proposed by the researchers towards handwritten Indian character recognition and many recognition systems for isolated handwritten numerals/characters are available in the literature. To get idea of the recognition results of different classifiers and to provide new benchmark for future research, in this paper a comparative study of Devnagari handwritten character recognition using twelve different classifiers and four sets of feature is presented. Projection distance, subspace method, linear discriminant function, support vector machines, modified quadratic discriminant function, mirror image learning, Euclidean distance, nearest neighbour, k-Nearest neighbour, modified projection distance, compound projection distance, and compound modified quadratic discriminant function are used as different classifiers. Feature sets used in the classifiers are computed based on curvature and gradient information obtained from binary as well as gray-scale images.
2014 22nd International Conference on Pattern Recognition, 2014
This work addresses the problem of creating a Bayesian Network based online semi-supervised handw... more This work addresses the problem of creating a Bayesian Network based online semi-supervised handwritten character recognisor, which learns continuously over time to make a adaptable recognisor. The proposed method makes learning possible from a continuous inflow of a potentially unlimited amount of data without the requirement for storage. It highlights the use of unlabelled data for boosting the accuracy, especially when labelled data is scarce and expensive unlike unlabelled data. An algorithm is introduced to perform semi-supervised learning based on the combination of novel online ensemble of the Randomized Bayesian network classifiers and a novel online variant of the Expectation Maximization (EM) algorithm. We make use of a novel varying weighting factor to modulate the contribution of unlabelled data. Proposed method was evaluated using online handwritten Tamil characters from the IWFHR 2006 competition dataset. The accuracy obtained was comparable to the state of the art batch learning methods like HMM and SVMs.
Uploads
Papers by Umapada Pal