schirrmacher
commited on
Upload folder using huggingface_hub
Browse files
README.md
CHANGED
@@ -7,6 +7,8 @@ pretty_name: Human Segmentation Dataset
|
|
7 |
|
8 |
# Human Segmentation Dataset
|
9 |
|
|
|
|
|
10 |
This dataset was created **for developing the best fully open-source background remover** of images with humans.
|
11 |
The dataset was crafted with [LayerDiffuse](https://github.com/layerdiffusion/LayerDiffuse), a Stable Diffusion extension for generating transparent images.
|
12 |
|
@@ -15,19 +17,10 @@ Since Stable Diffusion is not perfect, the dataset contains images with flaws. S
|
|
15 |
|
16 |
The resulting model will be similar to [RMBG-1.4](https://huggingface.co/briaai/RMBG-1.4), but with open training data/process and commercially free to use.
|
17 |
|
18 |
-
I had some trouble with the Hugging Face file upload. You can find the data here: [Google Drive](https://drive.google.com/drive/folders/1K1lK6nSoaQ7PLta-bcfol3XSGZA1b9nt?usp=drive_link).
|
19 |
-
|
20 |
The dataset contains transparent images of humans (`/humans`) which are randomly combined with backgrounds (`/backgrounds`). Then the ground truth (`/gt`) for segmentation was computed based on the transparent images. The results are written to a training and validation dataset.
|
21 |
|
22 |
I created more than 5.000 images with people and more than 5.000 diverse backgrounds.
|
23 |
|
24 |
-
# Examples
|
25 |
-
|
26 |
-
Here you can see an augmented image and the resulting ground truth:
|
27 |
-
|
28 |
-
![](example_image.png)
|
29 |
-
![](example_ground_truth.png)
|
30 |
-
|
31 |
# Create Training Dataset
|
32 |
|
33 |
The following script creates augmented training and validation data.
|
@@ -38,10 +31,19 @@ Notice: download the dataset from [Google Drive](https://drive.google.com/drive/
|
|
38 |
./create_dataset.sh
|
39 |
```
|
40 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
41 |
# Support
|
42 |
|
43 |
If you identify weaknesses in the data, please contact me.
|
44 |
|
|
|
|
|
45 |
# Changelog
|
46 |
|
47 |
- Added more diverse backgrounds (natural landscapes, streets, houses)
|
|
|
7 |
|
8 |
# Human Segmentation Dataset
|
9 |
|
10 |
+
[>>> Download Here <<<](https://drive.google.com/drive/folders/1K1lK6nSoaQ7PLta-bcfol3XSGZA1b9nt?usp=drive_link).
|
11 |
+
|
12 |
This dataset was created **for developing the best fully open-source background remover** of images with humans.
|
13 |
The dataset was crafted with [LayerDiffuse](https://github.com/layerdiffusion/LayerDiffuse), a Stable Diffusion extension for generating transparent images.
|
14 |
|
|
|
17 |
|
18 |
The resulting model will be similar to [RMBG-1.4](https://huggingface.co/briaai/RMBG-1.4), but with open training data/process and commercially free to use.
|
19 |
|
|
|
|
|
20 |
The dataset contains transparent images of humans (`/humans`) which are randomly combined with backgrounds (`/backgrounds`). Then the ground truth (`/gt`) for segmentation was computed based on the transparent images. The results are written to a training and validation dataset.
|
21 |
|
22 |
I created more than 5.000 images with people and more than 5.000 diverse backgrounds.
|
23 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
24 |
# Create Training Dataset
|
25 |
|
26 |
The following script creates augmented training and validation data.
|
|
|
31 |
./create_dataset.sh
|
32 |
```
|
33 |
|
34 |
+
# Examples
|
35 |
+
|
36 |
+
Here you can see an augmented image and the resulting ground truth:
|
37 |
+
|
38 |
+
![](example_image.png)
|
39 |
+
![](example_ground_truth.png)
|
40 |
+
|
41 |
# Support
|
42 |
|
43 |
If you identify weaknesses in the data, please contact me.
|
44 |
|
45 |
+
I had some trouble with the Hugging Face file upload. You can find the data here: [Google Drive](https://drive.google.com/drive/folders/1K1lK6nSoaQ7PLta-bcfol3XSGZA1b9nt?usp=drive_link).
|
46 |
+
|
47 |
# Changelog
|
48 |
|
49 |
- Added more diverse backgrounds (natural landscapes, streets, houses)
|