A novel Adaptive Shape Support Window based cost aggregation method for local stereo matching

Cost aggregation is the most important step in a local stereo matching algorithm. In this paper, a novel cost aggregation method based on Adaptive Shape Support Window (ASSW) is proposed. There are two main parts: At first we construct a local support skeleton anchored each pixel with four varying arm lengths decided on color similarity, as a result, the support window integral of multiple horizontal segments spanned by pixels in the vertical neighboring is established; then we utilize extended implementation of guided filtering to aggregate cost volume within the ASSW, which has better edge-preserving smoothing property than bilateral filter independent of the filtering kernel size. In this way, the proposed method can effectively reduce the number of bad pixels locating in the incorrect depth regions through finding optimal support window with an arbitrary shape and size adaptively. When the proposed cost aggregation method is implemented together with the remaining steps, the overall local stereo matching algorithm achieves more outstanding matching performance compared with other existing algorithms based on cost aggregation strategy on the Middlebury stereo benchmark, especially in depth discontinuities and piecewise smooth regions.