Most of depth up-sampling algorithms are based on the consistent hypothesis, i.e. the object boundaries in the colour image are consistent with depth discontinuity regions in the depth map. However, the hypothesis is not always correct. Under the combined guidance of high-resolution (HR) depth edge map and HR colour image gradient map, a simple and efficient depth up-sampler is presented. Firstly, the consistent regions are distinguished from the other regions and more accurate depth edge points are found. Then, the initial upsampled depth map from traditional bilinear interpolation is refined by an effective depth-assignment scheme. Extensive experiments demonstrate that the proposed method outperforms conventional interpolation algorithms and some other edge-based depth up-sampling methods.

An efficient depth map generation method is presented for static scenes with moving objects. Firstly, static background scene is reconstructed. Depth map of the reconstructed static background scene is extracted by linear perspective. Then, moving objects are segmented precisely. Depth values are assigned to the segmented moving objects according to their positions in the static scene. Finally, the depth values of the static background scene and the moving objects are integrated into one depth map. Experimental results show that the proposed method can generate smooth and reliable depth maps.

No abstract

No abstract

This paper presents a novel depth map generation method based on geometric information. Our method divides a 2d image into two parts-the foreground and the background. Through extracting the predominant lines and vanishing point of the background, the background depth map is determined. Then we use this depth map as a 'scalar' to measure the depth value of the foreground object. Finally, the depth map of the foreground and the background are integrated into one depth map by a depth fusion algorithm.