방송공학회논문지 2009년 제14권 제4호 397 a) 서울벤처정보대학원대학교 컴퓨터응용기술학과 Dept. of Computer Science & Application Seoul Univ. of Venture & Information b) 호서대학교 벤처전문대학원 IT 응용기술학과 Dept. of IT Application Hoseo Graduate School of Venture Hoseo University c) 경북대학교 컴퓨터공학과 Dept. of Computer Engineering Kyungpook National University ‡교신저자 : 김구진(kujinkim@knu.ac.kr) ※This research is supported by Foundation of Ubiquitous Computing and Networking project(UCN), the Ministry of Knowledge Economy (MKE) 21st Centry Frontier R&D Program in Korea and a result of subproject UCN (09C1-T3-12T). ․접수일(2009년3월13일),수정일(2009년6월30일),게재확정일(2009년7월20일) Retrieval of Identical Clothing Images Based on Non-Static Color Histogram Analysis Yoo-Joo Choi a) , Nam-Mee Moon b) , and Ku-Jin Kim c)‡ Abstract In this paper, we present a non-static color histogram method to retrieve clothing images that are similar to a query clothing. Given clothing area, our method automatically extracts major colors by using the octree-based quantization approach[16]. Then, a color palette that is composed of the major colors is generated. The feature of each clothing, which can be either a query or a database clothing image, is represented as a color histogram based on its color palette. We define the match color bins between two possibly different color palettes, and unify the color palettes by merging or deleting some color bins if necessary. The similarity between two histograms is measured by using the weighted Euclidean distance between the match color bins, where the weight is derived from the frequency of each bin. We compare our method with previous histogram matching methods through experiments. Compared to HSV cumulative histogram-based approach, our method improves the retrieval precision by 13.7 % with less number of color bins. Keywords: color quantization, color identification, histogram, clothing identification Ⅰ. Introduction The recent moderns have been exposed to a lot of multi- media data in the daily life. Not only image data but also various kinds of sound and movie data related to the every- day life have been dramatically increased. Therefore, the effective and interactive content-based retrieval methods for a large multimedia database have been actively re- searched in recent years [1-4] . Most of content-based re- trieval systems for an image database analyze color and texture information to define the image features and to dis- criminate the similarity of different images. Especially, the color analysis is a basic core technology for the robust im- age retrieval. Color analysis technologies have been also applied to the smart surveillance system to robustly track pedestrians from CCD camera images captured in real-time. In the multiple camera environment, the clothing colors and posi- tions of the pedestrians for each camera are analyzed and compared with the pedestrian information of the different camera in order to discriminate the same pedestrian [5-6] . Especially, the clothing color of the pedestrian is a primary analysis factor for person identification in the multiple camera environment that doesn’t include the overlapping 특집논문-09-14-4-02