Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking
Authors: Mingzhan Yang, Guangxin Han, Bin Yan, Wenhua Zhang, Jinqing Qi, Huchuan Lu, Dong Wang
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments Experimental Setting Datasets We evaluated our design on various MOT benchmarks, including Dance Track (Sun et al. 2022), MOT20 (Dendorfer et al. 2020) and MOT17 (Milan et al. 2016). |
| Researcher Affiliation | Collaboration | Mingzhan Yang1,2*, Guangxin Han1*, Bin Yan1, Wenhua Zhang1, Jinqing Qi1, Huchuan Lu1, Dong Wang1 1 Dalian University of Technology 2 Shenzhen Tvt Digital Technology Co., Ltd |
| Pseudocode | No | The paper describes its algorithm framework in text but does not include a structured pseudocode block or a clearly labeled algorithm. |
| Open Source Code | Yes | The code and models are available at https://github.com/ymzis69/Hybrid SORT. |
| Open Datasets | Yes | Datasets We evaluated our design on various MOT benchmarks, including Dance Track (Sun et al. 2022), MOT20 (Dendorfer et al. 2020) and MOT17 (Milan et al. 2016). |
| Dataset Splits | Yes | The MOT17 validation set follows a widely adopted convention (Zhou, Koltun, and Kr ahenb uhl 2020), where the train set is split into halves for training and validation. |
| Hardware Specification | Yes | The hardware is a single V100 GPU with Intel Xeon(R) Silver 4214R CPU @ 2.40GHz. |
| Software Dependencies | No | The paper mentions models like YOLOX and Bo T and concepts like FP16-precision but does not provide specific version numbers for key software dependencies or libraries (e.g., Python, PyTorch, CUDA versions). |
| Experiment Setup | Yes | The weight hyper-parameter of the confidence cost matrix in the first and second association stages are 1.5 and 1.0 on Dance Track, 1.0 and 1.0 on other benchmarks. The weight of ROCM cost is 0.2, the same as OCM in OC-SORT. The Io U threshold to reject a match is set to 0.15 on Dance Track, and 0.25 on other benchmarks. |