Simplified High-Performance Cost Aggregation for Stereo Matching

Applying edge preservation filters for cost aggregation has been a leading technique in generating dense disparity maps. However, traditional approaches usually require intensive calculations, and their design parameters must be tuned different scenarios to obtain the best performance. This paper shows that simple texture-independent approach can achieve similar high The proposed algorithm is equivalent sequence of matrix multiplications involving two weighting matrices primary matching cost. Notably, are constant image pairs with same resolution. For higher accuracy, we integrate multi-scale scheme fully exploit spatial distribution textures pairs. resultant hybrid efficient accurate enough surpass most existing stereo matching. performance verified by extensive simulation results using Middlebury (3rd Edition) benchmark database.