Adaptive resistive network for stereo depth estimation

Starting from a resistive network that performs, at each node, the convolution between an image and a Gabor function of any phase, the author presents a resistive network able to extract phase information from an image. An adaptation mechanism is used to find, for each node, the Gabor function that best matches the image at the node. This network is useful to evaluate the disparity between two stereo images for depth estimation