색역 범위 확장을 통한 컬러 영상의 색상 다채화

Abstract

This thesis presents a novel color-to-colorful conversion algorithm which reproduces a morecolorful image via image gamut coverage expansion. In this thesis, we first analyzes the statistics ofcolorful image databases to obtain the features of colorful images. then, we propose a set of criteriafor colorful image generation based on statistics of colorful image databases, human color perceptionstudies, and the dark channel prior. With the criteria, we develop an optimization frameworkthat can be easily solved as a linear system. Experimental results and user study demonstrate thatthe proposed method produce vivid, vibrant, colorful results for a variety of color images.