Fast Minimax Path-Based Joint Depth Interpolation

Longquan Dai     Feihu Zhang     Xing Mei     Xiaopeng Zhang

Abstract

We propose a fast minimax path-based depth interpolation method. The algorithm computes for each target pixel varying contributions from reliable depth seeds, and weighted averaging is used to interpolate missing depths. Compared with state-of-the-art joint geodesic upsampling method which selects the nearest seeds to interpolate missing depths with complexity, our method does not need to limit the number of seeds to and reduces the computational complexity to . In addition, the minimax path chooses a path with the smallest maximum immediate pairwise pixel difference on it, so it tends to preserve sharp depth discontinuities better. In contrast to the results of previous depth upsampling algorithms, our approach can provide accurate depths with fewer artifacts

Downloads

``Fast Minimax Path-Based Joint Depth Interpolation.''  Longquan Dai, Feihu Zhang, Xing Mei, Xiaopeng Zhang.  IEEE Signal Processing Letters 22.5 (2015): 623 - 627.    [paper]

Our Method

Experiments

Reference

[1] Q. Yang, R. Yang, J. Davis, and D. Nister, “Spatial-depth super resolution for range images,” in CVPR, 2007.

[2] D. Ferstl, C. Reinbacher, R. Ranftl, M. Ruether, and H. Bischof, “Image guided depth upsampling using anisotropic total generalized variation,” in IEEE Int. Conf. Computer Vision, Dec. 2013, pp. 993–100.