Feature Extraction Technique for Neural Network Based Pattern Recognition
Journal Title: International Journal on Computer Science and Engineering - Year 2012, Vol 4, Issue 3
Abstract
In this work, an attempt is made to extract minimum number of features to represent the pattern used as inputs for Feed Forward Back Propagation Neural Network (FFBPNN). The binary image of a pattern stored in the frame is partitioned into square regions. A feature from each region is computed by the density and co-ordinate distance of 1s.pixels. The neural network is trained with the extracted features and Root Mean Square Error (RMSE) obtained in the training process is used as performance indicator to stop the FFBPNN learning. Tested the proposed feature extraction and classification algorithms on the handwritten numeral database and found very good classification recognition rate.
Authors and Affiliations
Ashoka H. N. , Manjaiah D. H. , Rabindranath Bera
Keywords
Related Articles
- EP ID EP86577
- DOI -
- Views 149
- Downloads 0
Computational Approaches for determination of Most Probable RNA Secondary Structure Using Different Thermodynamics Parameters
ABSTRACT : Many bioinformatics studies require the analysis of RNA structures. More specifically, extensive work is done to elaborate efficient algorithms able to predict the 2-D folding structures of RNA. The core of RN...
Identity Recognizing using Iris Scan with multiple frames of Video
The field of iris biometrics has been significant research over the last decade. At this point, iris capture has become a main stream technology with wide acceptance. A general iris recognition system works with four dif...
Modern Trends Used In Operating Systems For High Speed Computing Applications
Operating system researches traditionally consist of adding new functions to the operating system in other words inventing and evaluating new methods for performing functions. Operating systems are the single most comple...
Performance Analysis of Lifetime Maximizing Trees for Data Aggregation in Wireless Sensor Networks
In this paper the performance of the proposed DLMT and CLMT algorithms are compared. These proposed algorithms tend to extend the node lifetime. Decentralized lifetime maximizing tree(DLMT) features in nodes with higher...
Efficient New Design and Verification of Sign-Digit-Adder for Two Symmetric Redundant Radix-4 Numbers
The carry propagation during the addition operation limits the speed of arithmetic operation. In Digital arithmetic, several approaches were proposed to provide the best implementation for addition operation of two opera...