Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Humans
|
| 2 |
+
|
| 3 |
+
This dataset was created _for developing the best fully open-source background remover_ of images with humans.
|
| 4 |
+
|
| 5 |
+
The dataset was crafted with
|
| 6 |
+
[LayerDiffuse](https://github.com/layerdiffusion/LayerDiffuse), a Stable Diffusion extension for generating transparent images.
|
| 7 |
+
|
| 8 |
+
It covers a diverse set of humans: various skin tones, clothes, hair styles etc.
|
| 9 |
+
Since Stable Diffusion is not perfect, the dataset contains images with flaws. Still the dataset is good enough for training background remover models.
|
| 10 |
+
|
| 11 |
+
The dataset contains transparent images of humans (`/humans`) which were randomly combined with backgrounds (`/backgrounds`).
|
| 12 |
+
Then the groundtruth (`/gt`) for segmentation was computed based on the transparent images. The results are written to a training and validation dataset.
|
| 13 |
+
|
| 14 |
+
I created 4.558 images with people which I have augmented to a set of 9.116 images for training and 2.451 for validation with diverse backgrounds. The backgrounds are optimized for art exhibitions because this is where I want to apply the background remover.
|
| 15 |
+
|
| 16 |
+
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
|