Segmentation of camera-trap tiger images based on texture and color features

Abstract

In this paper we propose a method to obtain the region of interest in a camera-trap tiger image so that it can be used for identification of the tiger in later stage. We use a combination of texture and color features to discriminate the tiger from the background. Based on these features, a level set based active contour segmentation is performed. A series of morphological operation are then performed on the segmented result to obtain the desired region of interest. The experimental results show the effectiveness of the proposed algorithm.