PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume

CVPR 2018 Deqing SunXiaodong YangMing-Yu LiuJan Kautz

We present a compact but effective CNN model for optical flow, called PWC-Net. PWC-Net has been designed according to simple and well-established principles: pyramidal processing, warping, and the use of a cost volume... (read more)

PDF Abstract

Evaluation results from the paper


Task Dataset Model Metric name Metric value Global rank Compare
Dense Pixel Correspondence Estimation HPatches PWC-Net Viewpoint I AEPE 4.43 # 2
Dense Pixel Correspondence Estimation HPatches PWC-Net Viewpoint II AEPE 11.44 # 3
Dense Pixel Correspondence Estimation HPatches PWC-Net Viewpoint III AEPE 15.47 # 3
Dense Pixel Correspondence Estimation HPatches PWC-Net Viewpoint IV AEPE 20.17 # 3
Dense Pixel Correspondence Estimation HPatches PWC-Net Viewpoint V AEPE 28.30 # 3