Alpha Estimation in Natural Images

Many boundaries between objects in the world project onto curves in an image. However, boundaries involving natural objects (e.g., trees, hair, water, smoke) are often unworkable under this model because many pixels receive light from more than one object. We propose a technique for estimating alpha, the proportion in which two colors mix to produce a color at the boundary. The technique extends bluescreen mattingtobackgrounds that have almost arbitrary color distributions, though coarse knowledge of the boundary’s location is required. Results show a number of different objects moved from one image to another while maintaining naturalism.

• Type: Conference
• Year: 2000
• Where: CVPR
• Authors: Mark A. Ruzon, Carlo Tomasi
• Keywords: Arbitrary Color Distributions, CVPR 2000, Many Boundaries, Pixels Receive Light