A novel robust depth estimation method based on optimal region selection

摘要

In this paper, a novel robust depth estimation method based on optimal region selection is proposed with improved anti-noise capability and structural retention. In particular, this new scheme provides the practitioners with a better de-noising ability by means of improving the non-subsampled contourlet transform (NSCT) features. Moreover, an optimal region selection technique is developed to further suppress the noise in focus measure. In order to make the features more prominent, the derivatives of features along optical axis are normalized for weighting in optimal region selection process. Experimental results demonstrate that the proposed method has superiority on better anti-noise ability, higher structural retention performance, compared with the existing representative methods. (C) 2019 Elsevier Ltd. All rights reserved.