Fast Adaptive Support-Weight Stereo Matching Algorithm

Abstract

Adaptive support-weight (ASW) stereo matching algorithm is widely used in the field of three-dimensional (3D) reconstruction owing to its relatively high matching accuracy. However, since all the weight coefficients need to be calculated in the whole disparity range for each pixel, the algorithm is extremely time-consuming. To solve this problem, a fast ASW algorithm is proposed using twice aggregation. First, a novel weight coefficient which adapts cosine function to satisfy the weight distribution discipline is proposed to accomplish the first cost aggregation. Then, the disparity range is divided into several sub-ranges and local optimal disparities are selected from each of them. For each pixel, only the ASW at the location of local optimal disparities is calculated, and thus, the complexity of the algorithm is greatly reduced. Experimental results show that the proposed algorithm can reduce the amount of calculation by 70% and improve the matching accuracy by 6% for the 15 images on Middlebury Website on average.

Keywords

Stereo matching / Cost aggregation / Adaptive support-weight algorithm / Weight coefficient

References

[1] Tombari F, Gori F, Stefano LD (2011) Evaluation of stereo algorithms for 3D object recognition. 2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops) 990–997

[2] Scharstein D, Szeliski R. A taxonomy and evaluation of dense two-frame stereo correspondence algorithms. Int J Comput Vis, 2002, 47: 7-42.

[3] Kolmogorov V, Zabih R. Computing visual correspondence with occlusions using graph cuts. Eighth IEEE Int Conf Comput Vis, 2001, 2: 508-515.

[4] Banno A, Ikeuchi K. Disparity map refinement and 3D surface smoothing via directed anisotropic diffusion. Comput Vis Image Underst, 2011, 115: 611-619.

[5] Wang ZF, Zheng ZG (2008) A region based stereo matching algorithm using cooperative optimization. IEEE Conference on Computer Vision and Pattern Recognition CVPR 1–8

[6] Chen DM, Ardabilian M, Chen LM. A fast trilateral filter based adaptive support weight method for stereo matching. IEEE Trans Circuits Syst Video Technol, 2015, 25: 730-743.

[7] Yang QX. Stereo matching using tree filtering. IEEE Trans Pattern Anal Mach Intell, 2015, 37: 834-846.

[8] Rhemann C, Hosni A, Bleyer M et al (2011) Fast cost-volume filtering for visual correspondence and beyond., IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 3017–3024

[9] He KM, Sun J, Tang XO. Guided image filtering. IEEE Trans Pattern Anal Mach Intell, 2013, 35: 1397-1409.

[10] Zhang K, Lu JB, Lafruit G. Crossed-based local stereo matching using orthogonal integral images. IEEE Trans Circuits Syst Video Technol, 2009, 19: 1073-1079.

[11] Cigla C, Alatan AA (2011) Efficient edge-preserving stereo matching. IEEE International Conference on Computer Vision Workshops 696–699

[12] Yoon KJ, Kweon IS. Adaptive support-weight approach for correspondence search. IEEE Trans Pattern Anal Mach Intell, 2006, 28: 650-656.

[13] Sun X, Mei X, Jiao SH et al (2011) Stereo matching with reliable disparity propagation. International Conference on 3D Image, Modelling, Processing, Visualization and Transmission (3DIMPVT) 132–139

[14] Li H, Zhang XG, Sun Z. A line-based adaptive-weight matching algorithm using loopy belief propagation. Math Probl Eng, 2015, 2015: 297392.

[15] Middlebury stereo Evaluation-Version 3. http://vision.middlebury.edu/stereo/eval3/

[16] Tan P, Monasse P. Stereo disparity through cost aggregation with guided filter. Image Process Line, 2014, 4: 252-275.

[17] Mattoccia S, Giardino S, Gambini A (2009) Accurate and efficient cost aggregation strategy for stereo correspondence based on approximated joint bilateral filtering. Asian Conference on Computer Vision 371–380

[18] Zhang K, Li JY, Li YJ et al. (2012) Binary stereo matching. 21st International Conference on Pattern Recognition (ICPR) 356–359