Planar Prior Assisted PatchMatch Multi-View Stereo

Authors: Qingshan Xu, Wenbing Tao12516-12523

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results demonstrate that our method can efficiently recover the depth information of extremely low-textured areas, thus obtaining high complete 3D models and achieving state-of-the-art performance.
Researcher Affiliation Academia National Key Laboratory of Science and Technology on Multispectral Information Processing School of Artifical Intelligence and Automation, Huazhong University of Science and Technology, China {qingshanxu, wenbingtao}@hust.edu.cn
Pseudocode Yes The overall pipeline of our algorithm is summarized in Algorithm 1.
Open Source Code Yes Our code will be available at https://github.com/Ghi Xu/ACMP.
Open Datasets Yes We evaluate the effectiveness of our method on high-resolution multi-view stereo dataset of ETH3D benchmark (Sch ops et al. 2017).
Dataset Splits Yes The dataset is further split into training datasets and test datasets.
Hardware Specification Yes Our methods are implemented in C++ with CUDA and executed on a machine with two Intel E5-2630 CPUs and two GTX Titan X GPUs.
Software Dependencies No The paper mentions 'implemented in C++ with CUDA' but does not provide specific version numbers for these software components or any other libraries.
Experiment Setup Yes {ϵ, α, γ, λn, σ, η, λgeo, τgeo} = {0.1, 0.18, 0.5, 5 , 0.3, 0.9, 0.1, 5.0}. Besides, λd is adaptively set to one sixty-fourth of the depth interval of every reference image.