AutoRemover: Automatic Object Removal for Autonomous Driving Videos
Authors: Rong Zhang, Wei Li, Peng Wang, Chenye Guan, Jin Fang, Yuhang Song, Jinhui Yu, Baoquan Chen, Weiwei Xu, Ruigang Yang12853-12861
AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments show that our method outperforms other state-of-the-art (SOTA) object removal algorithms, reducing the RMSE by over 19%. Additionally, the paper includes a dedicated section '4 Experiments and Results' with quantitative comparisons (Table 1 and Table 2) and visual comparisons (Figure 7 and Figure 8). |
| Researcher Affiliation | Collaboration | The authors are affiliated with '1Zhejiang University, 2Baidu Research, Baidu Inc. 3University of Southern California, 4Peking University', which includes a mix of academic institutions (Zhejiang University, University of Southern California, Peking University) and an industry research lab (Baidu Research, Baidu Inc.). |
| Pseudocode | No | The paper describes its architecture and procedures using text and diagrams (e.g., Figure 3, Figure 4) but does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks. |
| Open Source Code | No | The paper states, 'This dataset will be released to the public with the paper,' referring to their shadow dataset, but it does not provide an explicit statement or link for the release of their method's source code. |
| Open Datasets | Yes | The paper explicitly states the use of a public dataset: 'In our paper, we use the Apollo Scape (Huang et al. 2018) for experiments.' It also announces the release of a new dataset: 'This dataset will be released to the public with the paper,' referring to their shadow annotated video AD dataset. |
| Dataset Splits | No | The paper mentions 'Total 16373 samples of training sequence are generated' and the phrase 'validation' in the context of a metric ('TWE') but does not specify a distinct validation dataset split (e.g., as a percentage or specific number of samples) for hyperparameter tuning or model selection. |
| Hardware Specification | Yes | The paper explicitly states the hardware used for experiments: 'All the experiments are implemented with Tensorflow & Paddle Paddle and performed on 4 NVIDIA Tesla P40.' |
| Software Dependencies | No | The paper mentions the software frameworks used: 'All the experiments are implemented with Tensorflow & Paddle Paddle.' However, it does not provide specific version numbers for these or any other software dependencies. |
| Experiment Setup | Yes | The paper provides specific experimental setup details, including hyperparameters and training configurations: 'The network is trained with an Adam optimizer for 210k iterations, whose learning rate is 0.0001 and batch size is 8.' It also describes data preprocessing steps: 'During the training, the images are bottom cropped to 562 226 and then randomly cropped to 384 192 around the generated holes.' |