METHOD AND APPARATUS FOR CATEGORY-BASED PHOTO CLUSTERING USING PHOTOGRAPHIC REGION TEMPLATES OF DIGITAL PHOTO

摘要

A method and a system for clustering photos based on a category by using a section division template are provided to categorize semantic concept included in the photo by effectively dividing photo data into each section, extracting the semantic concept of each section, and efficiently merging the semantic concept of each section. A section divider(110) divides the photo into the sections by using the section template. A sectional semantic concept modeler(120) models the sectional semantic concept included in the photo by each divided section. A dominant concept extractor(130) extracts the dominant concepts of each section through modeling. A histogram generator(140) generates a histogram of the dominant concepts. A category determiner(150) determines the category of the photo from the histogram.