Digit Recognition using Fixed Feature Maps |
Author(s): |
| Venkat Giridhar Mareedu , MVSR Engineering College; Shrinath Raju, MVSR Engineering College; Apurva Chintakunta, MVSR Engineering College |
Keywords: |
| Digit Recognition, CNN, Feature Maps |
Abstract |
|
Digit Recognition is a well-researched and a well-documented field. But the practical applications have been limited due to the error rates and the impact of every single wrong prediction of digit on the output process. This paper aims at designing a recognition model with a specific application using fixed feature maps. This paper does not implement a feature map extractor but with the input digit image taken in a fixed format it uses a CNN with pre-defined positions of feature maps to predict digits with highest possible accuracy rate. |
Other Details |
|
Paper ID: IJSRDV5I60296 Published in: Volume : 5, Issue : 6 Publication Date: 01/09/2017 Page(s): 758-759 |
Article Preview |
|
|
|
|
