Handwritten Form Processing
Bandhan V1, K Neetish Bhat2, Karthik C Borkar3, Mamtha HR4
1Bandhan V, ISE department, PESIT, Bangalore (Karnataka), India.
2K Neetish Bhat, ISE department, PESIT, Bangalore (Karnataka), India.
3Karthik C Borkar, ISE department, PESIT, Bangalore (Karnataka), India.
4Dr. Mamtha HR, ISE department, PESIT, Bangalore (Karnataka), India.
Manuscript received on 15 June 2015 | Revised Manuscript received on 25 June 2015 | Manuscript Published on 30 June 2015 | PP: 24-27 | Volume-4 Issue-5, June 2015 | Retrieval Number: E3996064515/15©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Analysis of document images for information extraction has become vital in the modern day. These days so much variety of information is being conventionally stored on paper. For better storage and accurate processing, the paper is being converted into electronic form. This involves a lot of processing of documents using image processing techniques and other computer vision concepts. Pre-Processing techniques like Gaussian Blur, Otsu Thresholding, Median Filter and morphological operations are adopted to increase accuracy of recognition. Based on contours each fields of form are segmented. Character segmentation is done based on bounding box. MNSIT SD-19 database is used for training of characters. SVM and k-NN techniques are used for classification. Our implementation was tried for 10 requisition for certificate forms. Out of 10 forms 8 forms was correctly generated. So the accuracy of result is found to be 80%.
Keywords: Object Character Recognition; Pre-Processing; Segmentation; Classification; Post-Processing;
Scope of the Article: Classification