AI Paper Atlas

Research map / Point clouds / Shape and surface completion

Deep Point Set Resampling via Gradient Fields

Haolan Chen, Bi'an Du, Shitong Luo, Wei Hu

TPAMI 2022 · Point clouds › Shape and surface completion · approach: Point cloud upsampling and normals · 45 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add cosine annealing (continuous model) to resampling (ablation)Ours · PUNet denoising 10K, 1% noiseCD (x1e-4): 2.489 -> 2.353Improved
Add cosine annealing (continuous model) (ablation)Ours · PUNet denoising 10K, 3% noiseCD (x1e-4): 4.185 -> 4.075Improved
Add cosine annealing (continuous model) (ablation)Ours · PUNet denoising 10K, 2% noiseP2M (x1e-4): 0.812 -> 0.734Improved
Add regularization (GLR/RGLR) to resampling (ablation)Ours · PUNet denoisingCDImproved
Use reweighted GLR instead of GLR (ablation)Ours · PUNet denoisingCDImproved
Add gradient-field resampling on top of coarse generator at 4x (ablation)Ours-Gen · Point cloud upsampling 4xCD (x1e-4): 2.769 -> 2.324Improved
Add gradient-field resampling on top of coarse generator at 8x (ablation)Ours-Gen · Point cloud upsampling 8xCD (x1e-4): 2.161 -> 1.532Improved
Add gradient-field resampling on top of coarse generator at 16x (ablation)Ours-Gen · Point cloud upsampling 16xCD (x1e-4): 2.132 -> 0.840Improved
Use generator-based initialization instead of noise perturbationOurs-Naive · Point cloud upsampling 4xCD (x1e-4): 2.901 -> 2.324Improved
Use one-pass Ours-Naive instead of progressive MPU at 8xMPU · Point cloud upsampling 8xP2M (x1e-5): 1.828 -> 1.577Improved
Use Ours-Naive instead of MPU at 4xMPU · Point cloud upsampling 4xCD (x1e-4): 2.348 -> 2.901Hurt
Use Ours instead of Score at 3% noise, 10K PUScore · PUNet denoising 10K, 3% noiseCD (x1e-4): 4.708 -> 4.075Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy