Datasets:
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# TFW: Annotated Thermal Faces in the Wild Dataset
|
2 |
+
|
3 |
+
## Dataset Summary
|
4 |
+
|
5 |
+
The TFW dataset comprises thermal images captured in three distinct environments: controlled indoor (c-indoor), semi-controlled indoor (s-indoor), and uncontrolled outdoor (u-outdoor). The c-indoor subset utilizes images from the SpeakingFaces dataset. The s-indoor and u-outdoor subsets were collected using a FLIR T540 thermal camera (464x348 pixels, 7.5–14 μm wave-band, 24° field of view, iron color palette). Each image is annotated with face bounding boxes and five facial landmarks (centers of both eyes, tip of nose, outer corners of mouth).
|
6 |
+
|
7 |
+
## Dataset Structure
|
8 |
+
|
9 |
+
The dataset is organized into three environment subsets:
|
10 |
+
|
11 |
+
* **c-indoor:** 142 subjects, 5,112 images, 5,112 labeled faces, visual pairs available.
|
12 |
+
* **s-indoor:** 9 subjects, 780 images, 1,748 labeled faces, visual pairs available.
|
13 |
+
* **u-outdoor:** 15 subjects, 4,090 images, 9,649 labeled faces, visual pairs not available.
|
14 |
+
|
15 |
+
A combined subset includes all environments.
|
16 |
+
|
17 |
+
|
18 |
+
## Data Examples
|
19 |
+
|
20 |
+
[Example image showing annotations](https://github.com/IS2AI/TFW/blob/main/figures/example.png)
|
21 |
+
|
22 |
+
## Dataset Download
|
23 |
+
|
24 |
+
The dataset can be downloaded from [Google Drive](https://drive.google.com/file/d/1q8jfnDdCe3e-YOnoL5PIpSlkt0vamPMV/view?usp=sharing) or via a request to [this link](https://issai.nu.edu.kz/tfw-annotated-thermal-faces-in-the-wild-dataset/).
|
25 |
+
|
26 |
+
|
27 |
+
## Pre-trained Models
|
28 |
+
|
29 |
+
Pre-trained YOLOv5 and YOLO5Face thermal face detection models are available for download. Details on the models can be found on [Google Drive](https://drive.google.com/drive/folders/10ToqjavIlk5bj63zV4xgibE1L-jTEzxN?usp=sharing) (YOLOv5) and [Google Drive](https://drive.google.com/drive/folders/12ub57wP1hZ4tL2WH7TrUpmbvXXIdi3NU?usp=sharing) (YOLO5Face). Example results are shown below.
|
30 |
+
|
31 |
+
[Example image showing YOLOv5 detection results](https://github.com/IS2AI/TFW/blob/main/figures/yolov5.png)
|
32 |
+
|
33 |
+
[Example image showing YOLO5Face detection results](https://github.com/IS2AI/TFW/blob/main/figures/yolov5_face.png)
|
34 |
+
|
35 |
+
[Demo GIF](https://github.com/IS2AI/TFW/blob/main/figures/demo.gif)
|
36 |
+
|
37 |
+
|
38 |
+
## Contact
|
39 |
+
|
40 |
+
For any questions or issues, please contact the dataset creators. Further information can be found in the associated paper.
|