Pattern Recognition using Normalized Feature Vectors Analysis
Year: 2016
Type: article
Abstract: Objective: To study pattern recognition and retrieval in machine vision system application. Methods/Analysis: Regular and irregular pattern recognition algorithm based on sorting of radii from centre of mass is used and other tools such as Matlab, Neural networks. Findings: The radii from centre of mass to contour of the pattern are computed and sorted in descending order. Few top radii are taken for recognition of the given pattern. As the radii are sorted in descending order, therefore, if the pattern is orientated at any angle, the top order radii are same. This enables the pattern recognition at any orientation. Further, the radii are normalized with respect to their mean radius to make them size invariant. In addition, area, perimeter and euler number are also computed for enhancing the uniqueness degree in features vector set. Keywords: CBIR, Extraction, Normalization, NN
Authors: Pranab Garg, Dinesh Kumar
Institution: Guru Kashi University
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W2478939852
