Rethinking Rotation Invariance with Point Cloud Registration
Authors: Jianhui Yu, Chaoyi Zhang, Weidong Cai
AAAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results on 3D shape classification, part segmentation, and retrieval tasks prove the feasibility of our work. |
| Researcher Affiliation | Academia | Jianhui Yu, Chaoyi Zhang, Weidong Cai School of Computer Science, University of Sydney, Australia {jianhui.yu, chaoyi.zhang, tom.cai}@sydney.edu.au |
| Pseudocode | No | The paper describes its methodology using textual explanations and mathematical equations, but it does not include any clearly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | Our project page is released at: https://rotation3d.github.io/. |
| Open Datasets | Yes | We first examine the model performance on the synthetic Model Net40 (Wu et al. 2015) dataset. ... Experiments are also conducted on a real-scanned dataset. Scan Object NN (Uy et al. 2019) is a commonly used benchmark... We use Shape Net Part (Yi et al. 2016) for evaluation... We further conduct 3D shape retrieval experiments on Shape Net Core55 (Chang et al. 2015)... |
| Dataset Splits | Yes | We combine the training and validation sets and validate our method on the testing set following the training policy of (Esteves et al. 2018). |
| Hardware Specification | No | The paper does not provide specific details about the hardware (e.g., GPU models, CPU types, memory) used for running the experiments. |
| Software Dependencies | No | The paper refers to deep learning architectures and modules (e.g., Point Net++, Transformer) but does not list specific software dependencies (e.g., Python version, PyTorch/TensorFlow version, CUDA version) with their version numbers. |
| Experiment Setup | Yes | Hyper-parameters for training follow the same as (Guo et al. 2021), except that points are downsampled in the order of (1024, 512, 128) with feature dimensions of (3, 128, 256). ... The training strategy is the same as the classification task except that the training epoch number is 300. |