Fig. 11From: Robust stereo matching with trinary cross color census and triple image-based refinementsResults in Middlebury 2015 stereo datasets. a Color image. b Ground truth. c Estimated disparity mapBack to article page