HF-NeuS: Improved Surface Reconstruction Using High-Frequency Details

Authors: Yiqun Wang, Ivan Skorokhodov, Peter Wonka

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 4 Experiments, Table 1: Quantitative results on the DTU dataset.
Researcher Affiliation Academia Yiqun Wang KAUST Ivan Skorokhodov KAUST Peter Wonka KAUST
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code Yes Code available at https://github.com/yiqun-wang/HFS.
Open Datasets Yes We conduct experiments on the DTU dataset [12]. We further choose 9 challenging scenes from other datasets: 6 scenes from the Ne RF-synthetic dataset [19] and 3 scenes from Blended MVS [31](CC-4 License).
Dataset Splits No The paper mentions using standard datasets like DTU, Ne RF-synthetic, and Blended MVS and following previous work's evaluation protocols, but it does not explicitly provide specific training, validation, and test dataset split percentages or counts.
Hardware Specification Yes We use Adam with learning rate 5e-4 for the network training using NVIDIA TITAN A100 40GB graphics cards.
Software Dependencies No The paper mentions using MLPs and the Adam optimizer but does not provide specific version numbers for any software libraries or dependencies used for implementation.
Experiment Setup Yes We use MLPs to model two signed distance functions fb and fd. Each MLP consists of 8 layers. We use Adam with learning rate 5e-4 for the network training... For adaptive sampling, we first uniformly sample 64 points on the ray... We set L = 16 for the parameter of the frequency band of positional encoding.