Gamut mapping for motion picture

Abstract

New technologies in capturing and displaying images with extended color gamut and new standards for wide color gamut color encoding enable a new market of extended-color-gamut content (cinema, television, video on demand, games, electronic documents). This paper addresses the scenario of introducing extended-color-gamut motion picture production, post-production and distribution. One technical issue is the management and compatibility of different color gamuts. Manual (high-end) and automatic (consumer) color conversion tools are needed. For this purpose, we analyze in this paper different gamut mapping algorithms for motion picture type content. A psycho-physical evaluation framework is developed that is based on motion-picture related quality criteria aiming the preservation of the artistic intent and the optimal use of destination color gamut. We notably focus on gamut mapping algorithms that analyze geometrically both the source and the destination color gamut. A new method for automatic detection of the cusp of a color gamut is developed and used for cusp to cusp gamut mapping. The evaluation of several gamut mapping algorithms shows that best results are obtained for cusp to cusp mapping. According to the motion picture evaluation framework, the cusp to cusp gamut mapping algorithm performs significantly better, notably for night scenes and animated content.