Depth-based segmentation — A review

Abstract

Computer Vision describes some of its typical assignments as motion analysis, scene reconstruction, image restoration and image segmentation. Enormous tasks could be demonstrated while segmenting an image. The goal might be stated as, “Extraction of object from background” or “Analysing sharp edges in the image”. The goal could be any, but the initial information needed to visualize the difference between different segments of image, should be accurately picked and evaluated. This paper talks about such an initial information i.e. depth value of image pixels and gives an insight of its importance in the field of Image segmentation. Moreover, the paper also describes some of the typical techniques to incur depth information in the process of segmentation, solely or combined with other pixel values like color.