RGBD-Dog: Predicting Canine Pose from RGBD Sensors

Figure 1 : We present a system to predict the skeleton pose of a dog from RGBD images. If the size and shape of the dog is unknown, an estimation is provided. Displayed here are frames [4,7,13, 18 ] of a Kinect sequence, showing 2D projection, 3D skeleton and skinned mesh as produced by the pipeline. All figures in this paper are most informative when viewed in colour.