Diffuse-Specular Separation and Depth Recovery from Image Sequences

European Conference on Computer Vision (ECCV) |

Publication

Specular reflections present difficulties for many areas of computer vision such as stereo and segmentation. To separate specular and diffuse reflection components, previous approaches generally require accurate segmentation, regionally uniform reflectance or structured lighting. To overcome these limiting assumptions, we propose a method based on color analysis and multibaseline stereo that simultaneously estimates the separation and the true depth of specular reflections. First, pixels with a specular component are detected by a novel form of color histogram differencing that utilizes the epipolar constraint. This process uses relevant data from all the stereo images for robustness, and addresses the problem of color occlusions. Based on the Lambertian model of diffuse reflectance, stereo correspondence is then employed to compute for specular pixels their corresponding diffuse components in other views. The results of color-based detection aid the stereo correspondence, which determines both separation and true depth of specular pixels. Our approach integrates color analysis and multibaseline stereo in a synergistic manner to yield accurate separation and depth, as demonstrated by our results on synthetic and real image sequences.