BLADE: Box-Level Supervised Amodal Segmentation through Directed Expansion
Authors: Zhaochen Liu, Zhixuan Li, Tingting Jiang
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments are conducted on several challenging datasets and the results show that our proposed method can outperform existing state-of-the-art methods with large margins. |
| Researcher Affiliation | Academia | 1National Engineering Research Center of Visual Technology, National Key Laboratory for Multimedia Information Processing, School of Computer Science, Peking University 2AI Innovation Center, School of Computer Science, Peking University 3School of Computer Science and Engineering, Nanyang Technological University 4National Biomedical Imaging Center, Peking University |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper does not provide a concrete link to source code or explicitly state that the code is publicly available. |
| Open Datasets | Yes | Our experiments are conducted on three challenging datasets, including Occluded Vehicles (Wang et al. 2020), KINS (Qi et al. 2019) and COCOA-cls (Follmann et al. 2019). |
| Dataset Splits | No | The paper describes the datasets and their characteristics (e.g., occlusion levels, number of objects) but does not provide specific train/validation/test splits by percentages, sample counts, or explicit splitting methodology. |
| Hardware Specification | Yes | The training is completed on 3 NVIDIA Ge Force RTX 2080Ti GPUs taking about 5 hours per time. |
| Software Dependencies | No | We implement our model based on Box Inst (Tian et al. 2021) and Detectron2 (Wu et al. 2019) on the Py Torch framework (Paszke et al. 2019). No specific version numbers are provided for these software components. |
| Experiment Setup | Yes | In the training, we use data at all occlusion levels and choose αa 1 = 2.0, αa 2 = 1.0, αa 3 = 1.0. We conduct 60000 iterations with a batch size of 6, during which we adopt a 3-stage learning rate of 0.01 in the first 40000 iterations, 0.001 in the 40000-54000 iterations, and 0.0001 in the 54000-60000 iterations. |