Chromaticity-based separation of reflection components in a single image

The separation of diffuse and specular reflection components, or equivalently specularity removal, is required in the fields of computer vision, object recognition and image synthesis. This paper proposes a simple and effective method to separate reflections in a color image based on the error analysis of chromaticity and appropriate selection of body color for each pixel. By solving the least-squares problem of the dichromatic reflection model, reflection separation is implemented on a single pixel level, without requiring image segmentation and even local interactions between neighboring pixels. Experimental evaluation indicates that the proposed method is effective and can deal with a wide variety of images. 2008 Elsevier Ltd. All rights reserved.