MMA: Multi-Camera Based Global Motion Averaging
Authors: Hainan Cui, Shuhan Shen490-498
AAAI 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments demonstrate that our algorithm achieves superior accuracy and robustness on various data sets compared to the state-of-the-art methods. |
| Researcher Affiliation | Collaboration | 1 National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing, China 2 School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China 3 CASIA-Sense Time Research Group, China |
| Pseudocode | No | The paper describes the proposed algorithms using mathematical formulations and descriptive text, but it does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks, nor does it present structured steps in a code-like format. |
| Open Source Code | No | The paper mentions that 'The code of BATA is provided in (Zhuang 2018)' which refers to a baseline method, and references third-party libraries like Ceres Solver and Theia, but it does not provide any link or explicit statement about the availability of the source code for their own proposed method (MMA). |
| Open Datasets | Yes | The stereo camera dataset comes from odometry benchmark of KITTI (Geiger, Lenz, and Urtasun 2012). |
| Dataset Splits | No | The paper mentions using datasets for experiments and for finding optimal parameters (ablation study for 'K'), but it does not provide specific details on training, validation, or test dataset splits (e.g., percentages, sample counts, or explicit splitting methodology) for reproduction. |
| Hardware Specification | No | The paper states 'All datasets are collected sequentially and run on a same computer with 256GB memory' and 'feature matching is run on GPU', but it does not provide specific details about the CPU model, GPU model, or other detailed hardware specifications. |
| Software Dependencies | No | The paper mentions the use of 'root-SIFT' for feature detection, 'Ceres Solver' for bundle adjustment, and refers to 'Theia' for baseline implementations, but it does not provide specific version numbers for these software components or any other libraries/languages used. |
| Experiment Setup | Yes | In our work, the parameter K in the translation averaging is set to 8. ... To guarantee the robustness of scene reconstruction, five datasets containing loops are used to find the optimal parameter. Fig. 4 shows the performance of our system under different parameter settings. ... Hence, based on this discovery, the parameter K is set to 8, which means that in our multi-camera based translation averaging, each image is constrained only by the best 8 edges connected to it. |