Experimental comparison of various image segmentation techniques
- 발행기관 서강대학교 일반대학원
- 지도교수 김종락
- 발행년도 2018
- 학위수여년월 2018. 2
- 학위명 석사
- 학과 및 전공 일반대학원 수학과
- 실제URI http://www.dcollection.net/handler/sogang/000000062817
- 본문언어 영어
- 저작권 서강대학교 논문은 저작권보호를 받습니다.
초록/요약
One of the most popular areas in machine learning is image analysis. Image analysis is divided into image representation, image segmentation and motion segmentation. In this paper, we consider image segmentation. We compare the similarities of various techniques with the original image. We used the Berkely image database & our personal photographs. The image segmentation technique was implemented using MATLAB. Our results show that when quad-tree segmentation, k-means segmentation, and MS were applied to clearly distinguished images, they have high similarity. When threshold based segmentation was applied to images with many colors or ambiguous objects, they have high similarity. This implies that we know which image segmentation method is more useful for a given image.
more