An Approach for Feature Extraction of Alpha-Numeric by Using Snakes and Principal Component Analysis for Its Recognition

Hits: 3206
Research areas:
Year:
2013
Type of Publication:
Article
Keywords:
Feature Extraction, Snake Algorithm, Principal Component Analysis, Character Recognition, Classification
Authors:
Hitesh Rajput; Tanmoy Som; Soumitra Kar
Journal:
IJAIM
Volume:
2
Number:
2
Pages:
31-36
Month:
September
Abstract:
A Snake is an active contour for representing object contours. Snakes are deformable splines (smooth curve segments) placed in a potential field which translate and deform to reduce their potential energy. Traditionally snake algorithms are often used to find edges in the grey level images, by according low potentials to areas of high contrast so that the snake seeks to match its contours to high contrast edges. In this paper we have shown the steps for image normalization and devised an approach to find the large scale features of alpha-numeral texts using snakes in combination of Principal Component Analysis. Feature extraction helps in alpha-numeral recognition.
Full text: IJAIM_166_Final.pdf

Indexed By