DC-PCN: Point Cloud Completion Network with Dual-Codebook Guided Quantization
Authors: Qiuxia Wu, Haiyang Huang, Kunming Su, Zhiyong Wang, Kun Hu
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments on the PCN, Shape Net_Part, and Shape Net34 datasets demonstrate the state-of-the-art performance of our method. |
| Researcher Affiliation | Academia | 1South China University of Technology 2The University of Sydney EMAIL, EMAIL, EMAIL |
| Pseudocode | No | The paper describes the methodology in the 'Methodology' section using text and a diagram (Figure 2), but does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | Code https://github.com/tthrvfd/dcpcn. |
| Open Datasets | Yes | Datasets. To demonstrate the effectiveness of the proposed DC-PCN, we conducted the evaluation on three widely used datasets. PCN, Shape Net_Part, and Shape Net34. PCN (Yuan et al. 2018) consists of 28,974 shapes for training and 1,200 shapes for testing from 8 categories. ... Shape Net_Part contains 14,473 shapes and was split into 11,705 shapes for training and 2,768 for testing. ... Shape Net34. To evaluate the method s generalizability, following the experimental setup in (Yu et al. 2023)... Performance on KITTI Dataset To further demonstrate the generalizability of DC-PCN, we conducted evaluation on the KITTI dataset as shown in Table 3... |
| Dataset Splits | Yes | PCN (Yuan et al. 2018) consists of 28,974 shapes for training and 1,200 shapes for testing from 8 categories. ... Shape Net_Part contains 14,473 shapes and was split into 11,705 shapes for training and 2,768 for testing. ... Shape Net34. ... we partitioned the Shape Net dataset into two distinct subsets: 34 visible categories and 21 unseen categories. Within the visible categories, a subset of 100 objects per category was randomly selected to comprise the visible test set, ensuring the remaining objects were utilized for training purposes.For the unseen categories, a comprehensive collection of 2,305 objects was designated as the test set. |
| Hardware Specification | Yes | All experiments were conducted utilizing two RTX 2080Ti GPUs |
| Software Dependencies | No | The paper does not explicitly provide specific software dependencies with version numbers. |
| Experiment Setup | No | The paper discusses evaluation metrics and datasets but does not explicitly provide specific experimental setup details such as hyperparameter values (e.g., learning rate, batch size, number of epochs) or optimizer settings in the main text. |