Training deep neural networks for stereo vision

We present a method for extracting depth information from a rectified image pair. Our approach focuses on the first stage of many stereo algorithms: the matching cost computation. We approach the problem by learning a similarity measure on small image patches using a convolutional neural network. Training is carried out in a supervised manner by constructing a binary classification data set with examples of similar and dissimilar pairs of patches. We examine two network architectures for learning a similarity measure on image patches. The first architecture is faster than the second, but produces disparity maps that are slightly less accurate. In both cases, the input to the network is a pair of small image patches and the output is a measure of similarity between them. Both architectures contain a trainable feature extractor that represents each image patch with a feature vector. The similarity between patches is measured on the feature vectors instead of the raw image intensity values. The fast architecture uses a fixed similarity measure to compare the two feature vectors, while the accurate architecture attempts to learn a good similarity measure on feature vectors. The output of the convolutional neural network is used to initialize the stereo matching cost. A series of post-processing steps follow: cross-based cost aggregation, semiglobal matching, a left-right consistency check, subpixel enhancement, a median filter, and a bilateral filter. We evaluate our method on the KITTI 2012, KITTI 2015, and Middlebury stereo data sets and show that it outperforms other approaches on all three data sets