- The implementation for "Hyperspectral Video Tracking with Spectral-Spatial Fusion and Memory Enhancement".
- IEEE Transactions on Image Processing, 2025.
🏃Keep updating🏃:
- Trained models of SpectralTrack and SpectralTrack+ have been released.
- Training and testing codes of SpectralTrack and SpectralTrack+ have been released.
- Tracking results of SpectralTrack and SpectralTrack+ have been released.
| Benchmark | SpectralTrack (Pre/Suc) | SpectralTrack+ (Pre/Suc) |
|---|---|---|
| HOTC20 | 0.954 / 0.727 | 0.954 / 0.728 |
| NIR23 | 0.918 / 0.715 | 0.940 / 0.743 |
| RedNIR23 | 0.691 / 0.563 | 0.747 / 0.607 |
| VIS23 | 0.883 / 0.681 | 0.901 / 0.695 |
| NIR24 | 0.937 / 0.750 | 0.938 / 0.763 |
| RedNIR24 | 0.705 / 0.539 | 0.692 / 0.531 |
| VIS24 | 0.726 / 0.575 | 0.711 / 0.551 |
| MSSOT | 0.845 / 0.560 | 0.805 / 0.545 |
| MSVT | 0.975 / 0.748 | 0.963 / 0.737 |
git clone https://github.com/YZCU/SpectralTrack.git
- CUDA 11.8
- Python 3.9.18
- PyTorch 2.0.0
- Torchvision 0.15.0
- numpy 1.25.0
-
Training: Please download the hyperspectral training and testing sets: HOTC20, HOTC23, HOTC24, MSSOT, MSVT.
-
Fast Training: Download the pre-trained model of SpectralTrack and SpectralTrack+. Put it into
<pretrained_models>. -
Run
<tracking/0train_SpectralTrack.py>and<tracking/0train_SpectralTrack+.py>to train SpectralTrack and SpectralTrack+, respectively. -
The well-trained SpectralTrack model is put into
<output/train/yzcu/yzcu/yzcu_ep0030.pth.tar>. SpectralTrack+--><output/train/yzcu/yzcu+/yzcu_ep0030.pth.tar>. -
We have also released the well-trained SpectralTrack and SpectralTrack+ tracking models.
-
Testing: Run
<tracking/1test_SpectralTrack+.py>for testing, and results are saved in<output/results/yzcu/yzcu>.<tracking/1test_SpectralTrack+.py>--><output/results/yzcu/yzcu+>. -
Evaluating: Please download the evaluation benchmark Toolkit and vlfeat for more accurate evaluation.
-
Refer to the Hyperspectral Object Tracking Challenge for detailed evaluations.
-
Evaluation of the SpectralTrack and SpectralTrack+ tracker. Run
<tracker_benchmark_v1.0\perfPlot.m>
- If you find our work helpful in your research, kindly consider citing it. We appreciate your support！
@ARTICLE{11007172,
author={Chen, Yuzeng and Yuan, Qiangqiang and Xie, Hong and Tang, Yuqi and Xiao, Yi and He, Jiang and Guan, Renxiang and Liu, Xinwang and Zhang, Liangpei},
journal={IEEE Transactions on Image Processing},
title={Hyperspectral Video Tracking with Spectral-Spatial Fusion and Memory Enhancement},
year={2025},
volume={},
number={},
pages={1-1},
keywords={Feature extraction;Hyperspectral imaging;Photonic band gap;Foundation models;Visualization;Video tracking;Tracking;Training;Transformers;Imaging;Hyperspectral video tracking;Multi-modal video tracking;Parameter-efficient fine-tuning},
doi={10.1109/TIP.2025.3569479}}
- If you have any questions or suggestions, feel free to contact me.
- Email: yzchen1006@163.com
❤️ ❤️ We sincerely appreciate the insightful feedback provided by Editors and Reviewers. ❤️ ❤️