A Fast Algorithm for Image Geometric Feature Parameters Extraction

Abstract

How to extract the features quickly and correctly is the essential precondition for image identification and classification. As there existing large computation problems in the traditional feature extraction algorithms, this paper proposes a faster algorithm. Firstly, the image connected region is encoded by using the Vertex Chain Code to map as a closed area. Then a unified computational framework is developed to combine the Vertex Chain Code and Discrete Green, turning surface integral into line integral. Experimental results show that this algorithm is fast and effective, and can be used for special area.