In-Hand 3D Object Reconstruction from a Monocular RGB Video
Authors: Shijian Jiang, Qi Ye, Rengan Xie, Yuchi Huo, Xiang Li, Yang Zhou, Jiming Chen
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We evaluate our approach on HO3D and HOD datasets and demonstrate that it outperforms the state-of-the-art methods in terms of reconstruction surface quality, with an improvement of 52% on HO3D and 20% on HOD. |
| Researcher Affiliation | Collaboration | Shijian Jiang1, Qi Ye1,2*, Rengan Xie3, Yuchi Huo3,4, Xiang Li5, Yang Zhou5, Jiming Chen1 1College of Control Science and Engineering, Zhejiang University 2Key Lab of CS&AUS of Zhejiang Province 3State Key Lab of CAD&CG, Zhejiang University 4Zhejiang Lab 5OPPO US Research Center |
| Pseudocode | No | The paper describes the methodology using text and mathematical equations, but does not include explicit pseudocode blocks or algorithm listings. |
| Open Source Code | Yes | Project webpage: https://east-j.github.io/ihor. |
| Open Datasets | Yes | To evaluate our method, we perform the experiments on HO3D (Hampali et al. 2020) and HOD (Huang et al. 2022). |
| Dataset Splits | No | The paper mentions using "500 frames from each sequence of HO3D and all provided frames of HOD" for experiments and discusses a training process. However, it does not provide specific percentages or absolute counts for training, validation, and test dataset splits, nor does it refer to predefined standard splits for these datasets for reproducibility. |
| Hardware Specification | Yes | The training takes about 14 hours in total on a single NVIDIA RTX3090 GPU. |
| Software Dependencies | No | The paper mentions "Adam optimizer" and "Py Torch" but does not provide specific version numbers for these or any other software dependencies. |
| Experiment Setup | Yes | For training the model, we use Adam optimizer (Kingma and Ba 2014) with a learning rate of 5e 4 and sampled 1024 rays per batch for a total of 100k iterations. ... where λmask = 10, λeik = 0.1, λseg = 0.1, λcontact = 5 are set empirically. ... In our study, we empirically set τ = 0.001, β = 0.5. |