Contrastive Predictive Autoencoders for Dynamic Point Cloud Self-Supervised Learning
This addresses the problem of learning spatiotemporal representations from unlabeled point cloud data for applications like action recognition, though it is incremental as it builds on existing contrastive and generative paradigms.
The paper tackles self-supervised learning for point cloud sequences by introducing a method combining contrastive prediction and reconstruction, achieving performance comparable to supervised methods on action and gesture recognition benchmarks.
We present a new self-supervised paradigm on point cloud sequence understanding. Inspired by the discriminative and generative self-supervised methods, we design two tasks, namely point cloud sequence based Contrastive Prediction and Reconstruction (CPR), to collaboratively learn more comprehensive spatiotemporal representations. Specifically, dense point cloud segments are first input into an encoder to extract embeddings. All but the last ones are then aggregated by a context-aware autoregressor to make predictions for the last target segment. Towards the goal of modeling multi-granularity structures, local and global contrastive learning are performed between predictions and targets. To further improve the generalization of representations, the predictions are also utilized to reconstruct raw point cloud sequences by a decoder, where point cloud colorization is employed to discriminate against different frames. By combining classic contrast and reconstruction paradigms, it makes the learned representations with both global discrimination and local perception. We conduct experiments on four point cloud sequence benchmarks, and report the results on action recognition and gesture recognition under multiple experimental settings. The performances are comparable with supervised methods and show powerful transferability.