Integration of colour and affine invariant feature for multi-view depth video estimation. (17th August 2016)

• Record Type: Journal Article
• Authors: Zuo, Y.; An, P.; Shen, L.; Li, C.; Ma, R.
• Is Part Of: Imaging science journal. Volume 64:Number 6(2016)
• Page Start: 313
• Page End: 320
• Publication Date: 2016-08-17
• Subjects: Depth estimation -- Multi-view video plus depth (MVD) -- Affine invariant feature; Imaging systems -- Periodicals; 621.36705
• DOI: 10.1080/13682199.2016.1208969
• Languages: English
• ISSNs: 1368-2199

Abstract

Abstract : Although many depth map estimation methods have been developed, depth map estimation still has some difficulties in low-texture regions and occlusion. This paper presents a novel approach for multi-view depth video estimation which adopts an adaptive matching scheme for high-texture regions and low-texture regions, respectively. For low-texture regions, the proposed method integrates affine invariant features with colour to make the matching more robust. Furthermore, according to the smoothness assumption of depth maps, a refinement is then performed on the estimated depth maps. Experimental results show that the proposed method can achieve better or comparable performances than the state-of-the-art method in the category of local methods even with the less running time.