A two-level strategy for segmenting center of interest from pictures

摘要

Segmenting center of interests (COIs) from pictures is an important but highly challenging problem for researchers in computer vision and image processing. The capability of understanding the meanings of pictures by computers can lead to breakthroughs in a wide range of applications including Web image search and online picture-sharing communities. In this paper, a two-level strategy is presented, which consists of a rough segmentation stage and a fine segmentation stage. In the first level, a picture is partitioned into four regions by using a block clustering method based on color and texture features, and the COI within the picture is distinguished from the background according to the principles of photographic composition. This stage aims to determine the approximate region of the target. In the second level, a novel active contour model is established based on shape information and vector method, where the image energy is defined by a hue gradient and the external energy is generated from either a triangular inner force or a supplementary force. This stage tries to extract the boundary of the target accurately. Experimental results on photos downloaded from the Internet show the feasibility and effectiveness of the proposed method.