Fusion of Geometry and Color Information for Scene Segmentation

Scene segmentation is a well-known problem in computer vision traditionally tackled by exploiting only the color information from a single scene view. Recent hardware and software developments allow to estimate in real-time scene geometry and open the way for new scene segmentation approaches based on the fusion of both color and depth data. This paper follows this rationale and proposes a novel segmentation scheme where multidimensional vectors are used to jointly represent color and depth data and normalized cuts spectral clustering is applied to them in order to segment the scene. The critical issue of how to balance the two sources of information is solved by an automatic procedure based on an unsupervised metric for the segmentation quality. An extension of the proposed approach based on the exploitation of both images in stereo vision systems is also proposed. Different acquisition setups, like time-of-flight cameras, the Microsoft Kinect device and stereo vision systems have been used for the experimental validation. A comparison of the effectiveness of the different depth imaging systems for segmentation purposes is also presented. Experimental results show how the proposed algorithm outperforms scene segmentation algorithms based on geometry or color data alone and also other approaches that exploit both clues.