An acceleration of a graph cut segmentation with FPGA
Daichi Kobori, Tsutomu Maruyama · 2012
Image segmentation is one of the most important steps in image processing. The graph cut is an effective method for the image segmentation. For calculating the graph cut, the max-flow algorithm is widely used, but it requires long computation time. To execute the graph cut in real-time, the acceleration of max-flow algorithm with hardware is necessary. In this paper, we propose an implementation of a max-flow problem for the image segmentation on FPGA. In this system, the push-relabel method and the gap relabeling are used in order to achieve high performance on FPGA. The performance is 20-30 fps for standard benchmark images.