# Computing the Stereo Matching Cost with a Convolutional Neural Network Source: [Computing the Stereo Matching Cost with a Convolutional Neural Network]() Collection index: [llms.txt]() Data updated: 2025-10-25T20:14:30.423Z Authors: Jure \\v{Z}bontar, Yann LeCun Published: 2014-09-15T16:54:42.000Z ## Abstract We present a method for extracting depth information from a rectified image pair. We train a convolutional neural network to predict how well two image patches match and use it to compute the stereo matching cost. The cost is refined by cross-based cost aggregation and semiglobal matching, followed by a left-right consistency check to eliminate errors in the occluded regions. Our stereo method achieves an error rate of 2.61 % on the KITTI stereo dataset and is currently (August 2014) the top performing method on this dataset. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Stereo Matching on KITTI Stereo 2015 (test)](): Stereo Matching; KITTI Stereo 2015 (test); Error Rate (\> 3px): 6.34. - [Stereo Matching on Middlebury 2014 (test)](): Stereo Matching; Middlebury 2014 (test); Bad Pixel Rate (Thresh 2.0): 8.08. - [Stereo Matching on KITTI 2015 (test)](): Stereo Matching; KITTI 2015 (test); D1 Error (Overall): 3.89. - [Stereo Matching on KITTI 2012 (test)](): Stereo Matching; KITTI 2012 (test); Outlier Rate (3px, Noc): not recorded. ## Original sources and implementation [Original paper]() [PDF]()