DeepLabCut utilized the feature detectors (ResNets + readout layers) of one of the state-of-the-art algorithms for human pose estimation by Insafutdinov et al., called DeeperCut, which inspired the name for our toolbox (see references below). Since this time, the package has changed ...