SSTrack: Sample-interval Scheduling for Lightweight Visual Object Tracking
Authors: Yutong Kou, Shubo Lin, Liang Li, Bing Li, Weiming Hu, Jin Gao
IJCAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | With the collective effort from both perspectives, our method achieves State-of-the-Art CPUreal-time accuracy on 5 UAV-tracking benchmarks and 5 general object tracking benchmarks. |
| Researcher Affiliation | Collaboration | Yutong Kou1,2,3, Shubo Lin1,2,3, Liang Li5, Bing Li1,3,6, Weiming Hu1,2,3,4 and Jin Gao1,2,3 ... EMAIL, EMAIL, EMAIL |
| Pseudocode | No | The paper describes the methodology in prose and figures, but does not include any explicitly labeled 'Pseudocode' or 'Algorithm' blocks. |
| Open Source Code | Yes | Codes and models will be available at https://github.com/ Kou-99/SSTrack. |
| Open Datasets | Yes | We first compared our method with 6 State-of-the-Art (SOTA) CPU-real-time object trackers on 5 general object tracking datasets, namely La SOT [Fan et al., 2019], La SOText [Fan et al., 2021], GOT-10k [Huang et al., 2019], Tracking Net [Muller et al., 2018], NFS [Kiani Galoogahi et al., 2017]. ... An additional bin [0, 1) is added to hold samples from the image dataset COCO [Lin et al., 2014]. |
| Dataset Splits | Yes | GOT-10k requires trackers to be trained solely on the training split to verify generalization ability. We follow this protocol and the result is reported in Tab. 1. The result is averaged across videos from La SOT [Fan et al., 2019] training split. |
| Hardware Specification | Yes | The model is trained on 4 NVIDIA V100 GPU and the inference speed is measured on Intel Xeon Gold 6146 CPU. |
| Software Dependencies | No | The paper mentions specific baseline models and pre-trained weights (OSTrack [Ye et al., 2022], Vi T-Tiny [Dosovitskiy et al., 2020], D-MAE from [Wang et al., 2023; Gao et al., 2025]), but does not provide specific version numbers for any software, libraries, or frameworks (e.g., Python, PyTorch, CUDA versions). |
| Experiment Setup | Yes | For the interval-based sampler, we set the maximum sample interval M = 200 following OSTrack. We part M into 7 bins with equal spacing 30, which is the frame rate for the training video datasets. An additional bin [0, 1) is added to hold samples from the image dataset COCO [Lin et al., 2014]. The success-aware scheduler uses the average iou from the latest epoch as the success threshold and uses the schedule function with order 2. |