Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -9,13 +9,56 @@ dataset_info:
|
|
9 |
dtype: string
|
10 |
splits:
|
11 |
- name: train
|
12 |
-
num_bytes: 1749995
|
13 |
num_examples: 981
|
14 |
download_size: 1845563
|
15 |
-
dataset_size: 1749995
|
16 |
configs:
|
17 |
- config_name: default
|
18 |
data_files:
|
19 |
- split: train
|
20 |
path: data/train-*
|
|
|
21 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9 |
dtype: string
|
10 |
splits:
|
11 |
- name: train
|
12 |
+
num_bytes: 1749995
|
13 |
num_examples: 981
|
14 |
download_size: 1845563
|
15 |
+
dataset_size: 1749995
|
16 |
configs:
|
17 |
- config_name: default
|
18 |
data_files:
|
19 |
- split: train
|
20 |
path: data/train-*
|
21 |
+
license: unknown
|
22 |
---
|
23 |
+
|
24 |
+
# CK+ Dataset
|
25 |
+
|
26 |
+
## Overview
|
27 |
+
|
28 |
+
The CK+ dataset contains adapted data from the CK+ dataset, which includes 920 images of facial expressions. Each image has been preprocessed to 48x48 pixels, converted to grayscale, and cropped to focus on facial features using haarcascade_frontalface_default.
|
29 |
+
|
30 |
+
## Emotions
|
31 |
+
|
32 |
+
The dataset labels emotions based on the following indices:
|
33 |
+
|
34 |
+
- 0: Anger (45 samples)
|
35 |
+
- 1: Disgust (59 samples)
|
36 |
+
- 2: Fear (25 samples)
|
37 |
+
- 3: Happiness (69 samples)
|
38 |
+
- 4: Sadness (28 samples)
|
39 |
+
- 5: Surprise (83 samples)
|
40 |
+
- 6: Neutral (593 samples)
|
41 |
+
- 7: Contempt (18 samples)
|
42 |
+
|
43 |
+
## Usage Split
|
44 |
+
|
45 |
+
The dataset is split into three parts for different purposes:
|
46 |
+
|
47 |
+
- Training: 80%
|
48 |
+
- PublicTest: 10%
|
49 |
+
- PrivateTest: 10%
|
50 |
+
|
51 |
+
## Citation
|
52 |
+
|
53 |
+
If you use this dataset in your research or work, please cite it as follows:
|
54 |
+
|
55 |
+
```bibtex
|
56 |
+
@misc{yourcitationkey,
|
57 |
+
author = {Your Name},
|
58 |
+
title = {CK+ Dataset},
|
59 |
+
year = {Year of Publication},
|
60 |
+
publisher = {Hugging Face Dataset Hub},
|
61 |
+
journal = {Hugging Face Hub},
|
62 |
+
howpublished = {\url{https://huggingface.co/datasets/ckplus}},
|
63 |
+
note = {Accessed: YYYY-MM-DD}
|
64 |
+
}
|