Location Dependency in Video Prediction
This work addresses a specific limitation in video prediction for computer vision applications, but it appears incremental as it modifies existing architectures rather than introducing a new paradigm.
The authors tackled the problem of video prediction by addressing the spatial invariance of convolutional neural networks, which prevents modeling location-dependent patterns, and they proposed location-biased convolutional layers that significantly outperformed spatially invariant models.
Deep convolutional neural networks are used to address many computer vision problems, including video prediction. The task of video prediction requires analyzing the video frames, temporally and spatially, and constructing a model of how the environment evolves. Convolutional neural networks are spatially invariant, though, which prevents them from modeling location-dependent patterns. In this work, the authors propose location-biased convolutional layers to overcome this limitation. The effectiveness of location bias is evaluated on two architectures: Video Ladder Network (VLN) and Convolutional redictive Gating Pyramid (Conv-PGP). The results indicate that encoding location-dependent features is crucial for the task of video prediction. Our proposed methods significantly outperform spatially invariant models.