![]() ![]() In the first part, document image has been segmented into text and non-text regions using a texture-based analysis which separates the two different areas employing five Gray Level Co-occurrence Matrix (GLCM) features and structural analysis. Our complete system can be divided into two primary parts: Document image segmentation and Bangla character recognition. We present a holistic and pragmatic approach which is capable of efficiently handling documents with or without pictures and generating expected outcome. As a result, most of the currently available BCR methods fail to produce the desired output and almost all of them just outline the process instead of actually implementing it. Not any of the prevailing methods can handle documents with graphics, which is a grave inconvenience as a large portion of office documents, books and newspapers contain pictures. At present, several efficient character recognition systems exist for major languages and it is unfortunate that despite being one of the most popular languages of the world, till now there is no complete and efficacious Bangla character recognition (BCR) system for text extraction from document image. Character recognition is one of the eminent sectors of modern pattern recognition and machine learning. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |