Foreground and Background information based Word Spotting for Kannada Documents |
Author(s): |
| Somnath Biradar , Karnatak Arts, Science and Commerce College Bidar, India; Veershetty C, Karnatak Arts, Science and Commerce College Bidar, India; Prabha, Karnatak Arts, Science and Commerce College Bidar, India |
Keywords: |
| Word Retrieval, Morphology filters, Digital Image Processing (DIP), Image Retrieval (IR) |
Abstract |
|
In this paper, we proposed a method foreground and background information based word Spotting for Kannada documentsusing morphology filters. Morphological filters employed to capture global properties of the underlying image. An input document image is segmented into words and stored into a library. Then morphology filters are employed to represent the words of a large dataset as well as a query word. Then the cosine distance is used to measure the similarity between two words, based on it,the relevance of the word is estimated by generating distance ranks. Then correctly matched words are selectedat different distance thresholds such as97, 98 and 99 percent. Encouraging results are achieved in terms of the average precision rateat 79.20, average recall rate as 91.00 and F measure as84.57asa threshold 98%. |
Other Details |
|
Paper ID: NCACCP051 Published in: Conference 8 : NCACC-2016 Publication Date: 01/10/2016 Page(s): 258-262 |
Article Preview |
|
Download Article |
|
|
