Loading [a11y]/accessibility-menu.js
Fuzzy min-max neural network based translation, rotation and scale invariant character recognition using RTSI features | IEEE Conference Publication | IEEE Xplore

Fuzzy min-max neural network based translation, rotation and scale invariant character recognition using RTSI features


Abstract:

This paper proposes a character recognition system that is invariant to translation, rotation and scale. The system has two main sections namely, feature extraction and r...Show More

Abstract:

This paper proposes a character recognition system that is invariant to translation, rotation and scale. The system has two main sections namely, feature extraction and recognition. The feature extraction is carried out using RTSI (rotation, translation, and scale invariant) features. The main advantage of this feature vector is that it doesn't require the normalization of character. These features are very simple to implement as compared to other methods. The fuzzy min-max neural network (FMNN) is used in the recognition phase. The four dimensional RTSI feature vector consists of normalized moment of inertia, centroid length ratio, centroid sum, and normalized centroid sum. The character recognition systems is tested on 26 uppercase typed English capital letters with various fonts such as Ariel Unicode, Ariel Narrow, Microsoft scan serif and hand written characters.
Date of Conference: 16-16 September 2004
Date Added to IEEE Xplore: 30 November 2004
Print ISBN:0-7695-2216-5
Conference Location: Wuhan, China

References

References is not available for this document.