On high-computational disparity acquisition of HD image using CUDA implementation for stereo matching

In this paper, we propose disparity map acquisition of high definition image using CUDA with new search range estimation algorithm. Our proposed search range estimation method uses SSD with processing in gray scale image only. To perform high speed computation of disparity map, we used shared memory in CUDA, and utilize SSD as a matching algorithm to apply for a 9 by 9 window to 1240×1110 image, with using search range length of 250. Our experimental results showed that our method achieved 31fps on the average using NVIDIA GeForce 770.