Fast Learning of Signed Distance Functions from Noisy Point Clouds via Noise to Noise Mapping
This addresses a key bottleneck in 3D computer vision for applications like reconstruction and denoising, though it is incremental as it builds on existing SDF learning methods.
The paper tackles the problem of learning signed distance functions (SDFs) from noisy point clouds without ground truth supervision, achieving state-of-the-art results in surface reconstruction, denoising, and upsampling with training convergence within one minute.
Learning signed distance functions (SDFs) from point clouds is an important task in 3D computer vision. However, without ground truth signed distances, point normals or clean point clouds, current methods still struggle from learning SDFs from noisy point clouds. To overcome this challenge, we propose to learn SDFs via a noise to noise mapping, which does not require any clean point cloud or ground truth supervision. Our novelty lies in the noise to noise mapping which can infer a highly accurate SDF of a single object or scene from its multiple or even single noisy observations. We achieve this by a novel loss which enables statistical reasoning on point clouds and maintains geometric consistency although point clouds are irregular, unordered and have no point correspondence among noisy observations. To accelerate training, we use multi-resolution hash encodings implemented in CUDA in our framework, which reduces our training time by a factor of ten, achieving convergence within one minute. We further introduce a novel schema to improve multi-view reconstruction by estimating SDFs as a prior. Our evaluations under widely-used benchmarks demonstrate our superiority over the state-of-the-art methods in surface reconstruction from point clouds or multi-view images, point cloud denoising and upsampling.