ETRI-Knowledge Sharing Plaform

KOREAN
논문 검색
Type SCI
Year ~ Keyword

Detail

Conference Paper 내용기반 이미지 검색을 위한 회전 변형에 강인한 에지 히스토그램 알고리즘
Cited - time in scopus Share share facebook twitter linkedin kakaostory
Authors
김영근, 김선태, 오원근, 김의정
Issue Date
2009-11
Citation
대한전자공학회 종합 학술 대회 (추계) 2009, pp.373-376
Publisher
대한전자공학회
Language
Korean
Type
Conference Paper
Abstract
Content-base image retrieval is popular method for searching similar and copied images. It extracts the unique features, such as color, shape and texture, from images and compares each other with the specific distance. Edge histogram descriptor which is the one of the texture-based methods uses the distribution of the edge in an image with 5 directions. Edge feature in this descriptor has strong point for many modifications but weak in rotations. In this paper, we proposed the rotation-tolerant edge histogram for personal photo album service as a image retrieval application. From the experimental results, success ratio of edge histogram descriptor in rotation modifications is highly enhanced without degradation in other modifications.
KSP Keywords
Content-Base Image Retrieval, Edge features, Edge histogram descriptor(EHD), Personal photo, Success ratio, Texture-based, photo album