Activities of the Unit
ICDAR WML 2019
ML and its Application to PR
The major areas of on-going research in the Unit include Optical Character Recognition, Natural Language Processing, Pattern Recognition, Image Processing and Soft Computing tools.
The CVPR unit is proud of a pioneering development of an Optical Character Recognition (OCR) system for the combined script recognition of Bangla and Devnagari (Hindi), the most popular scripts of the Indian sub-continent. The system performs document tilt correction, script line, word and character segmentation, character recognition and finally, error correction. Its performance has been found to be about 98%. Work on Oriya OCR and OCR for Bangla hand-written character recognition is in progress.
A major area of interest of this unit is the general problem of digital document analysis. Novel algorithms for skew estimation and correction have been developed. A system for the automatic detection and processing of mathematical expressions from documents is being designed. The issues of automatic layout analysis of both printed and handwritten documents involving segmentation, information and feature extraction, data compression, processing and understanding line drawings are under study while a system for the automatic extraction of meta-information from document images has already been developed. A Bangla programming interface to the Turbo-C compiler has been implemented which has been found to work well for programs of moderate complexity. Another work in progress is language data compression, both from hypertext and images of documents in Indian scripts.
There have been pioneering developments in the area of Natural Language Processing of Indian languages. A powerful Bangla spell-checker for automatic detection and correction of spelling errors in computer stored documents has been developed and extended to the spell checking of Hindi documents. Among other important developments are a talking dictionary and talking editor for Bangla text, a morphological processor for Bangla and a computer based Bangla speech synthesizer. One of the novel objectives is to develop computer aids for the handicapped such as the deaf and dumb as well as the blind. Among other studies, statistical analysis of a large corpus of Bangla text, analysis of human spelling error patterns as well as analysis of onomatopoeic Bangla words have been carried out. Work on developing a tri-lingual computer dictionary (Hindi, Bangla, English) is in full swing.
In the area of bio-medical image processing, the processing of histo-pathological images acquired with a confocal laser scanning microscope for the recognition and grading of diseased cells has been studied at length. New algorithms for 3-D segmentation of these images and feature extraction have been proposed. Bayesian approaches are also being tested on MRI images for brain modelling.
Among soft computing tools, considerable research has been devoted to studying and developing applications of neural networks and genetic algorithms. A modified back propagation algorithm for better convergence and performance has been proposed and tested successfully for real-life problems like segmentation of natural textures and detection of linear structures in satellite images. New neural network models for the purpose of shape analysis in 2 and 3 dimensions have been developed. These models, viz. the dynamic self-organizing neural network and topology adaptive self-organizing neural network, overcome the inherent limitations of the existing models such as Kohonen's self-organizing feature map. A new self-organizing neural network has also been developed for the solution of a well known optimization problem, viz. finding the smallest circle enclosing a given set of planar points. Shape matching and shape detection in a point pattern, symmetry analysis and the k-centre problems are the issues being explored using genetic algorithms.
Some new clustering approaches are proposed where new definitions like nearest centroid neigbourhood have been brought in. The definition of neighbourhood is generalized to fuzzy sets and concepts like convex hull and circular hull have been extended to the digital domain both in crisp and fuzzy cases.
Work has recently started on automatic information retrieval of documents in Indian languages, in pictorial as well as hypertext form. Encouraging results have been obtained on a Bangla database of small size.
The Unit encourages its members to collaborate with various industrial as well as research and development organizations and actively engage themselves in both short-term and long-term projects. The list of ongoing projects funded by external agencies:
- Online Handwriting Recognition System for Indian Languages (OHWR) (An ongoing consortium-mode project funded by Ministry Of Communication And Information Technology, Govt. of India ).
- Development of Robust Document Analysis and Recognition System for Printed Devanagari and Bangla Scripts (An ongoing consortium-mode project funded by Ministry Of Communicat ion And Information Technology, Govt. of India).
- Cross Language Information Access Portal for Indian Languages (An ongoing consortium-mode project funded by Ministry Of Communication And Information Technology, Govt. of India).
- Online Handwritten Character Recognition using Character Sub-Strokes and Discriminative HMM Classifier (An ongoing project funded by HP Research Lab, India).
- Recognition of text in images of natural scenes (An ongoing International Joint Project in collaboration with Bristol University, UK funded by the Royal Society, UK).
Externally funded projects completed successfully in the recent past include:
- Indo-German Scientific Collaboration on Biomedical Image Processing (An ongoing project between ISI and GSF, Munich ).
- Indo-Japanese Collaboration on Natural Language Processing (An ongoing project between ISI and University of Foreign Studies, Tokyo).
- Development of Bilingual (Bangla and Devnagari) OCR system (An ongoing project funded by the Department of Science and Technology, Government of India).
- Development of a spell checker and morphological processor in an Indian language with speech output for the blind (An ongoing project sponsored by the Department of Electronics, Government of India).
- Speech synthesis (Malayalam) application for the blind and speech output for machine translation (An ongoing project together with the International School of Dravidian Linguistics, Thiruvananthapuram, sponsored by the Department of Science and Technology, Government of India).
Industrial collaboration and technical assistance
In addition to advanced level research, the Unit is interested in industrial collaboration, turnkey projects and consultancy. The areas in which the Unit can take up projects include, among others:
- Handwriting recognition (both on-line and off-line).
- Automatic form processing.
- OCR and document processing software.
- Text to speech conversion and multi-media environment.
- Picture data compression.
- Information retrieval from data in electronic and hypertext form.
- Natural language translation.
- Word processing application software such as spell checker.
- Biomedical image processing and applications.
Teaching, Training and Manpower Development Activities
In addition to their research activities, the faculty members of the Unit regularly participate in the teaching of various undergraduate, graduate and certain diploma courses offered by the Indian Statistical Institute, such as Pattern Recognition, Image Processing, Document Processing and retrieval, Signal Processing, Programming and Numerical Techniques, Compiler Construction, Operating System, Computer Architecture, Networks, Information Technology , Computational Geometry, Discrete Mathematics etc. A significant number of students of B.Tech, M.Tech, and MCA curriculum of various Universities and Institutes receive intensive training of duration 1 to 6 months every year under the guidance of the faculty members of the Unit.
In the past, its faculty members have organized various specialized training programs in collaboration with various Government and private organizations.
During the last few years the faculty of CVPR Unit have organized a number of Workshops at different Institutes of North-East India such as NEHU (Shillong), NERIST (Itanagar), Manipur University (Imphal), Mizoram University (Aizawl), SMIT (Gangtok) and SIT (Siliguri).
Forthcoming Research Activities
In keeping with the spirit of the Indian Statistical Institute there is a constant endeavour among the members of the unit to reach new heights in their research. Consequently, there are extensive plans for further research and development in all the areas mentioned earlier. In the area of document processing, these include study of the issues of document information extraction, text structuring, text query answering and data compression for Indian language documents. There are also plans of improving the performance of the current bilingual OCR system (Bangla and Devnagari) for printed multi-font and multi-sized text and further extension of the system to handle documents of poor visual quality. In the field of Biomedical Image Processing, the development of an automatic grading system for cancer based on 3-D images and preliminary work on functional brain mapping based on 4-D brain images will be considered. The novel self-organizing neural network model already developed will be further adapted for shape analysis while the issue of improving the speed of genetic algorithms by a cascaded approach will be explored.
In the area of Natural Language Processing, future plans include the development of a robust lexicon parser in Bangla, an automatic read-out system for Bangla texts and various studies like linguistic analysis of Bangla word forms, stylo-statistical analysis of Bangla sentences, etc. Other plans include the development and implementation of a voice and pitch detection program, intonation analysis and its incorporation in synthesized speech, incorporation of grapheme to phoneme conversion rules in the existing speech synthesis system, prosodic analysis of Bangla speech to improve the quality of synthesized Bangla speech and the study of Natural Language Processing techniques for application to high-precision information retrieval. Speech Corpus development for analysis and synthesis will be started soon. In addition to these, other issues that will be addressed are the development of theoretical methods and algorithms for basic problems of pattern recognition and image processing such as consistent set estimation from a point pattern, genetic algorithm based dot pattern analysis, convex hull detection of a digital binary object, and advancement of fuzzy clustering methods.
Students who have completed in the recent past
|Debasis Chaudhuri ||Some studies on density estimation and data clustering techniques|
| Nirupam Sarkar ||On texture image analysis using fractal geometry based features |
Supervisor: B.B. Chaudhuri
| Umapada Pal || On the development of an Optical Character Recognition (OCR) system for printed Bangla text |
Supervisor: B.B. Chaudhuri
| Anirban Ray Chaudhuri || Some studies on shape of dot patterns |
Supervisor: B.B. Chaudhuri
| P.S. Umesh Adiga || On quantitative evaluation of 3-D histopathological images obtained using confocal microscope |
Supervisor: B.B. Chaudhuri
| Amitava Datta || On some self-organizing models and their applications |
Supervisor: S.K. Parui
|Kirani Singh || On some self-organizing models and their applications |
Supervisor: S.K. Parui
|Ujjwal Bhattacharya || Some studies on recognition of handwritten characters of Bangla |
Co-Supervisor: S.K. Parui