A High-Dimensional Video Sequence Completion Method with Traffic Data Completion Generative Adversarial Networks

Figure 4

The completion results of single frame on Caltech pedestrian dataset. (a) Input/ground truth. (b) Output. (c) Details.