Probabilistic Shape Completion by Estimating Canonical Factors with Hierarchical VAE
This addresses the problem of generating complete 3D shapes from incomplete data for applications like robotics and computer vision, representing an incremental improvement over existing methods.
The paper tackles 3D shape completion from partial point clouds by proposing a method that estimates canonical factors of a feature volume using a hierarchical VAE, avoiding limitations of global latent codes or autoregressive approaches. It demonstrates effectiveness through quantitative and qualitative comparisons with state-of-the-art methods and ablation studies.
We propose a novel method for 3D shape completion from a partial observation of a point cloud. Existing methods either operate on a global latent code, which limits the expressiveness of their model, or autoregressively estimate the local features, which is highly computationally extensive. Instead, our method estimates the entire local feature field by a single feedforward network by formulating this problem as a tensor completion problem on the feature volume of the object. Due to the redundancy of local feature volumes, this tensor completion problem can be further reduced to estimating the canonical factors of the feature volume. A hierarchical variational autoencoder (VAE) with tiny MLPs is used to probabilistically estimate the canonical factors of the complete feature volume. The effectiveness of the proposed method is validated by comparing it with the state-of-the-art method quantitatively and qualitatively. Further ablation studies also show the need to adopt a hierarchical architecture to capture the multimodal distribution of possible shapes.