IRVAL

IRVAL is a high-resolution infrared video dataset for infrared video processing and spatial-temporal video super-resolution research.

Dataset Summary

This repository currently provides 8 infrared videos in .avi format.

According to our associated paper, IRVAL is a high-resolution infrared dataset comprising 108,512 video frames at a spatial resolution of 512×512. The data are collected using vanadium oxide (VOx) uncooled focal plane array detectors operating in the long-wave infrared (LWIR) band. The videos are captured from both vehicle-mounted and fixed surveillance platforms, covering real-world scenarios such as urban streets, vehicles, pedestrians, and roadside buildings.

Repository Structure

.
├── README.md
├── IRVAL/
│   └── videos/
│       ├── irval_seq01.avi
│       ├── irval_seq02.avi
│       ├── irval_seq03.avi
│       ├── irval_seq04.avi
│       ├── irval_seq05.avi
│       ├── irval_seq06.avi
│       ├── irval_seq07.avi
│       └── irval_seq08.avi
└── .gitattributes

Intended Use

This dataset is intended for research on:

• infrared video processing
• infrared video super-resolution
• spatial-temporal video super-resolution
• temporal consistency modeling

Notes

• This release currently contains 8 raw infrared videos.
• The current video path used by the dataset viewer is IRVAL/videos/*.avi.
• Users can generate task-specific LR/HR training pairs following their own protocol or the protocol described in the associated paper.
• This dataset is intended for research use only.

Citation

If you use this dataset, please cite:

@inproceedings{zhou2026thermal,
  title={Thermal Diffusion Matters: Infrared Spatial-Temporal Video Super-Resolution through Heat Conduction Priors},
  author={Mingxuan Zhou and Shuang Li and Yutang Zhang and Jing Geng and Yirui Shen and Jingxuan Kang and Fuzhen Zhuang and Shuigen Wang},
  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year={2026}
}