In this thesis, we propose a chroma keying system that automatically estimates the
alpha map and the reliable intrinsic color of foreground objects in front of solid background. Our system is designed to be capable of distinguishing the transparent
foreground from the re flective foreground and shaded background, thereby making
the artifacts of the composited image less conspicuous. Speci cally, we assume that
the transparent region tends to be with higher saturation and lightness compared with region re ecting background light. With this assumption, a threshold function (TF) on a saturation-lightness plane is de ned according to human visual experiments. The pixels with color mixed with the background light (conventional unknown pixels) are now further categorized into re ective pixels and transparent pixels according to TF. In this case, the re ective and the transparent regions are separated to improve the alpha matte quality. Furthermore, a new color representation model is proposed to estimate the intrinsic color of each pixel according to the global color distribution of the image. The underlying assumption of our proposed model is that all colors in a natural image can be approximated by a limited number of chrominance values (dominant colors). Speci cally, the color statistics are counted by 2D histogram analysis. Then, we approximate the color distribution by the sum of a set of Gaussian mixture functions (GMF), whose centroids are the dominant colors (Dc) of the image. By choosing colors around each Dc, the possible intrinsic colors for each pixel can be comprehensively and e fficiently selected. Considering the fast development of IP-based video broadcasting, we present a data hiding scheme that can protect the chroma keying results when the image/video data is recorded in the JPEG/H.264 format. According to our simulation, the proposed chroma keying system generates high quality composited images that are little a ected by re ecting, background shading, and intrinsic color missing.
Identifer | oai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/35271 |
Date | January 2016 |
Creators | Wang, Wenyi |
Contributors | Zhao, Jiying |
Publisher | Université d'Ottawa / University of Ottawa |
Source Sets | Université d’Ottawa |
Language | English |
Detected Language | English |
Type | Thesis |
Page generated in 0.0016 seconds