Group Contextual Encoding for 3D Point Clouds
Authors: Xu Liu, Chengtao Li, Jian Wang, Jingbo Wang, Boxin Shi, Xiaodong He
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate the effectiveness and generalizability of our method on three widely-studied 3D point cloud tasks. Experimental results have shown that the proposed method outperformed the Vote Net remarkably with 3 m AP on the benchmark of SUN-RGBD, with the metrics of m AP@0.25, and a much greater margin of 6.57 m AP on Scan Net with the metrics of m AP@0.5. |
| Researcher Affiliation | Collaboration | Xu Liu1,2 Chengtao Li3 Jian Wang4 Jingbo Wang5 Boxin Shi6 Xiaodong He2 1The University of Tokyo 2JD AI Research 3MIT 4Snap Inc. 5CUHK 6Peking University |
| Pseudocode | No | The paper describes procedures and uses equations but does not contain a formally labeled pseudocode or algorithm block. |
| Open Source Code | Yes | The source code3 has been released to facilitate the reproduction of our results. 3https://github.com/Asahi Liu/Point Detectron |
| Open Datasets | Yes | Dataset. SUN RGB-D [17] for 3D indoor scene understanding... Scan Net [4] provides a wider range of indoor scenes... |
| Dataset Splits | No | In our experiment, following [13] we split the training/testing set and report 3D detection performance on the 10 most common categories. ... We use the 1205 scans for training and 312 scans for testing, respectively. The paper does not explicitly mention a separate validation set or split details for it. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware (e.g., GPU models, CPU types, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not provide specific version numbers for any software dependencies, such as programming languages, libraries, or frameworks used for implementation. |
| Experiment Setup | Yes | We choose K = 8, C 3, G = 12 as the default setting in the following experiment. ... In experiments, we follow the same protocol in [13] and use the metrics, mean average precision (m AP), at Io U threshold of 0.25 for evaluation. |