전자어류도감을 위한 어류영상검색시스템에 관한 연구

Abstract

This paper propose the method of image segmentation and image retrieval for an Image-Retrieval System. Some image's object is similar to background in brightness, and if image segmentation is conducted by using existed algorithms in them, there will be a possibility of recognizing some object to background. The proposed image segmentation algorithm solves it by using local object extraction and object linking and the proposed image retrieval algorithm induces more accurate search by using efficient feature extraction and similarity function. The proposed feature extraction makes the head, body, and tail features through analysis object's length and area based on its major axis. The proposed similarity function decides priority and weight of features. The simulation results show that the proposed algorithm is superior to the conventional algorithms.