LightSAL: Lightweight Sign Agnostic Learning for Implicit Surface Representation
This work addresses efficiency bottlenecks in 3D shape reconstruction for applications requiring compact models and faster training, though it is incremental as it builds on existing Sign Agnostic Learning methods.
The paper tackles the problem of inefficient 3D shape modeling with implicit surface representations by proposing LightSAL, a lightweight convolutional architecture that reduces model size and training time while maintaining equivalent accuracy on the D-Faust dataset with 41k 3D scans.
Recently, several works have addressed modeling of 3D shapes using deep neural networks to learn implicit surface representations. Up to now, the majority of works have concentrated on reconstruction quality, paying little or no attention to model size or training time. This work proposes LightSAL, a novel deep convolutional architecture for learning 3D shapes; the proposed work concentrates on efficiency both in network training time and resulting model size. We build on the recent concept of Sign Agnostic Learning for training the proposed network, relying on signed distance fields, with unsigned distance as ground truth. In the experimental section of the paper, we demonstrate that the proposed architecture outperforms previous work in model size and number of required training iterations, while achieving equivalent accuracy. Experiments are based on the D-Faust dataset that contains 41k 3D scans of human shapes. The proposed model has been implemented in PyTorch.