Point-based Instance Completion with Scene Constraints

Authors: Wesley Khademi, Li Fuxin

ICLR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Through several experiments, we demonstrate that our method achieves improved fidelity to partial scans, higher completion quality, and greater plausibility over existing state-of-the-art methods.
Researcher Affiliation Academia Wesley Khademi & Li Fuxin Oregon State University EMAIL
Pseudocode No The paper describes the architecture and methods using textual descriptions and figures (e.g., Figure 2 and 3), but it does not include any clearly labeled pseudocode or algorithm blocks.
Open Source Code No The paper does not contain any explicit statement about releasing the source code for the methodology described, nor does it provide a link to a code repository.
Open Datasets No We introduce a new dataset called Scan WCF addressing the issues present in these existing datasets. On our newly proposed dataset we demonstrate that our approach outperforms existing works in terms of both partial reconstruction quality and completion quality while producing less collisions between predictions. The paper describes the Scan WCF dataset in detail and mentions it's composed of data from publicly available datasets, but it does not provide concrete access information (e.g., a direct link or DOI) for the Scan WCF dataset itself.
Dataset Splits Yes Our dataset contains 1202 indoor scenes based on Scan Net scenes, where 946 of the scenes are used for training and the other 246 scenes are reserved for testing. To increase the amount of training and test data, we generate 2 partial scans per scene using a different subset of camera poses from the Scan Net camera trajectory.
Hardware Specification Yes We train for 150 epochs with a batch size of 64 using 2 NVIDIA V100 GPUs, which takes approximately 4 days. Our scene completion model is then trained on the Scan WCF dataset, initialized from the weights of our pre-trained object completion model. For training, we use the same augmentation and optimization setup as used for pre-training. We train for 200 epochs on a single RTX 4090 GPU which takes about 3 days.
Software Dependencies No The paper mentions several software tools and frameworks like Adam (optimizer), Mask3D (Schult et al., 2023), NKSR (Huang et al., 2023), and PyTorch3D (Ravi et al., 2020), but it does not specify exact version numbers for these software dependencies.
Experiment Setup Yes For augmentations, we perform random rotations about the up-axis. For optimization, we use Adam with an initial learning rate of 1 10 4 and linearly decay by a factor of 0.98 every 2 epochs. We train for 150 epochs with a batch size of 64 using 2 NVIDIA V100 GPUs... We train for 200 epochs on a single RTX 4090 GPU.