# SF-TL54: A Thermal Facial Landmark Dataset with Visual Pairs

**Hugging Face Dataset Card**

**Summary**

The SF-TL54 dataset contains 2,556 thermal-visual image pairs of 142 subjects. Each pair features manually annotated face bounding boxes and 54 facial landmarks.  This dataset is derived from the larger SpeakingFaces dataset and is intended for research in thermal facial landmark detection.  The dataset includes images in both thermal and visual spectrums, offering a unique resource for multimodal facial analysis.

**Dataset Information**

| Attribute          | Value                                    |
|----------------------|-----------------------------------------|
| Name                | SF-TL54                                  |
| Description         | Thermal Facial Landmark Dataset with Visual Pairs |
| Number of Subjects  | 142                                      |
| Number of Images    | 2,556 (thermal-visual pairs)             |
| Annotations         | Face bounding boxes, 54 facial landmarks |
| Modality            | Thermal and Visual                        |
| Source Dataset      | SpeakingFaces                             |


**Example Image**

![Example Image](https://raw.githubusercontent.com/IS2AI/thermal-facial-landmarks-detection/main/figures/example.png)

**Facial Landmark Order**

![Landmark Configuration](https://raw.githubusercontent.com/IS2AI/thermal-facial-landmarks-detection/main/figures/land_conf.png)


**Citations**

```bibtex
@INPROCEEDINGS{9708901,
  author={Kuzdeuov, Askat and Koishigarina, Darina and Aubakirova, Dana and Abushakimova, Saniya and Varol, Huseyin Atakan},
  booktitle={2022 IEEE/SICE International Symposium on System Integration (SII)}, 
  title={SF-TL54: A Thermal Facial Landmark Dataset with Visual Pairs}, 
  year={2022},
  volume={},
  number={},
  pages={748-753},
  doi={10.1109/SII52469.2022.9708901}}

@Article{s21103465,
 AUTHOR = {Abdrakhmanova, Madina and Kuzdeuov, Askat and Jarju, Sheikh and Khassanov, Yerbolat and Lewis, Michael and Varol, Huseyin Atakan},
 TITLE = {SpeakingFaces: A Large-Scale Multimodal Dataset of Voice Commands with Visual and Thermal Video Streams},
 JOURNAL = {Sensors},
 VOLUME = {21},
 YEAR = {2021},
 NUMBER = {10},
 ARTICLE-NUMBER = {3465},
 URL = {https://www.mdpi.com/1424-8220/21/10/3465},
 ISSN = {1424-8220},
 DOI = {10.3390/s21103465}
}
```

**GitHub Repository**

[https://github.com/IS2AI/thermal-facial-landmarks-detection](https://github.com/IS2AI/thermal-facial-landmarks-detection)