Multi-view Consistent 3D Panoptic Scene Understanding
Authors: Xianzhu Liu, Xin Sun, Haozhe Xie, Zonglin Li, Ru Li, Shengping Zhang
AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experimental results demonstrate that the proposed MVC-PSU surpasses state-of-the-art methods on the Scan Net, Replica, and Hyper Sim datasets. |
| Researcher Affiliation | Academia | 1 Harbin Institute of Technology, Weihai, China 2 Nanyang Technological University, Singapore |
| Pseudocode | No | The paper describes the methods using mathematical formulas and descriptive text (e.g., equations for volume density, color, semantic logits, and loss functions) but does not present any structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not contain an explicit statement about releasing code or provide a link to a code repository. |
| Open Datasets | Yes | Following Panoptic Lifting (Siddiqui et al. 2023), we conduct experiments on three public datasets: Scan Net (Dai et al. 2017), Replica (Straub et al. 2019), and Hypersim (Roberts et al. 2021). |
| Dataset Splits | Yes | the available posed images in each dataset are divided into 75% for training views and 25% for testing views sampled in between. |
| Hardware Specification | No | The paper does not specify any hardware details like GPU models, CPU models, or memory used for experiments. |
| Software Dependencies | No | The paper mentions several models and networks like Tenso RF, Mask2Former, PDC-Net, and Dense Prediction Transformer (DPT), but does not provide specific version numbers for any of these software dependencies. |
| Experiment Setup | Yes | In the experiments, we set λdep = 1, while λsem, λint, and λseg are each set to 0.1. |