Swift-Mapping: Online Neural Implicit Dense Mapping in Urban Scenes

Authors: Ke Wu, Kaizhao Zhang, Mingzhe Gao, Jieru Zhao, Zhongxue Gan, Wenchao Ding

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experiments: We conduct experiments on both indoor and urban datasets. For indoor testing, we employ the synthetic Replica dataset (Straub et al. 2019) and the real-world Scan Net dataset (Dai et al. 2017). We also include datasets containing diverse urban scenes such as KITTI (Geiger et al. 2013), VKITTI2 (Cabon, Murray, and Humenberger 2020), and nu Scenes (Caesar et al. 2019).
Researcher Affiliation Academia Ke Wu1, Kaizhao Zhang2, Mingzhe Gao3, Jieru Zhao3, Zhongxue Gan1*, Wenchao Ding1* 1Academy for Engineering & Technology, Fudan University 2School of Future Technology, Harbin Institute of Technology 3Department of Computer Science and Engineering, Shanghai Jiao Tong University
Pseudocode No The paper does not contain structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide concrete access to source code for the methodology described in this paper.
Open Datasets Yes Datasets We conduct experiments on both indoor and urban datasets. For indoor testing, we employ the synthetic Replica dataset (Straub et al. 2019) and the real-world Scan Net dataset (Dai et al. 2017). We also include datasets containing diverse urban scenes such as KITTI (Geiger et al. 2013), VKITTI2 (Cabon, Murray, and Humenberger 2020), and nu Scenes (Caesar et al. 2019).
Dataset Splits No The paper does not provide specific dataset split information (exact percentages, sample counts, or detailed splitting methodology) needed to reproduce the data partitioning. It mentions following 'settings of NICE-SLAM' for the Replica dataset, but does not detail them in this paper.
Hardware Specification Yes For urban scenes, both Mip-Ne RF (Barron et al. 2022) and Instant NGP (M uller et al. 2022) take 10 minutes for training on a single NVIDIA RTX3090.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment.
Experiment Setup No The paper describes some general methodological choices and training aspects (e.g., L2 loss, optimizing certain parameters, 10 iterations in ablation) but does not provide specific experimental setup details such as concrete hyperparameter values (e.g., learning rate, batch size, optimizer settings) in the main text.