Add dataset card

README.md

@@ -1,3 +1,118 @@
 ---
-license: mit
+pretty_name: InScene
+language:
+  - en
+tags:
+  - face
+  - scene-restoration
+  - image-restoration
+  - diffusion
+  - synthetic
+  - facial-landmarks
+  - identity
+size_categories:
+  - 10K<n<100K
+task_categories:
+  - image-to-image
+  - text-to-image
 ---
+
+# InScene
+
+**InScene** is a synthetic dataset of full-body / in-the-wild scene images, each containing a person, paired with rich metadata (generation prompt, face bounding box, identity ID, and 68-point facial landmarks). It was created for the paper **Face2Scene: Using Facial Degradation as an Oracle for Diffusion-Based Scene Restoration** (CVPR 2026), where facial degradation is used as a supervisory signal ("oracle") for diffusion-based restoration of full scenes.
+
+## Why this dataset
+
+Faces are one of the most perceptually sensitive regions in an image and there exist strong, well-studied priors for facial quality. Face2Scene leverages this: by treating the face as an oracle for degradation, the model learns to restore the surrounding scene. To support this, InScene provides images in which a consistent set of **identities** appear across many diverse scenes, together with the localization metadata (face boxes and landmarks) needed to isolate and reason about the facial region.
+
+## Data generation
+
+Images were produced with the **InfiniteYou-based data-generation pipeline** available here:
+
+- 🔧 Pipeline: **https://github.com/amanwalia123/infiniteyoudatagen**
+
+Please refer to that repository for the full generation details (identity conditioning, prompt construction, sampling settings, etc.).
+
+## Dataset structure
+
+The dataset is organized into `train/` and `val/` splits. Each split is a flat directory of **sample folders**, one folder per generated image. Every sample folder contains exactly two files that share the sample's name:
+
+```
+train/
+  iden_00001_img_11_sample_321_repeat_1/
+    iden_00001_img_11_sample_321_repeat_1.png    # 1024x1024 RGB image
+    iden_00001_img_11_sample_321_repeat_1.json   # metadata for that image
+  ...
+val/
+  ...
+```
+
+The folder name encodes provenance: `iden_<identity>`, `img_<source-face-id>`, `sample_<sample-id>`, `repeat_<k>` (the same prompt/identity may be sampled multiple times).
+
+### Splits
+
+| Split   | Samples | Unique identities | Size  |
+|---------|--------:|------------------:|------:|
+| `train` |  11,260 |               870 |  19 GB |
+| `val`   |   1,329 |               100 |  3.4 GB |
+
+*(The `train` and `val` identity sets are disjoint.)*
+
+### Per-sample JSON schema
+
+| Field        | Type            | Description                                                                 |
+|--------------|-----------------|-----------------------------------------------------------------------------|
+| `image_name` | string          | Filename of the paired PNG.                                                 |
+| `prompt`     | string          | Structured scene prompt used for generation (subject, framing, background, detail/style). |
+| `image_size` | object          | `{ "width": 1024, "height": 1024 }`.                                        |
+| `face_bbox`  | list[int]       | Face bounding box `[x1, y1, x2, y2]` in pixel coordinates.                   |
+| `identity`   | string          | Zero-padded identity ID; the same ID denotes the same person across samples.|
+| `file_id`    | string          | Source face-image ID used to condition this identity.                       |
+| `landmark`   | list[[int,int]] | 68 facial landmark points `[x, y]` in pixel coordinates.                    |
+
+## Usage
+
+The dataset is a folder-of-folders (not a parquet/`imagefolder` layout), so the simplest way to use it is to download and glob the sample folders:
+
+```python
+import glob, json, os
+from PIL import Image
+from huggingface_hub import snapshot_download
+
+# Download just the train split (use allow_patterns="val/*" for val)
+local = snapshot_download(
+    "amir477/InScene",
+    repo_type="dataset",
+    allow_patterns="train/*",
+)
+
+samples = sorted(glob.glob(os.path.join(local, "train", "*")))
+folder = samples[0]
+name = os.path.basename(folder)
+
+image = Image.open(os.path.join(folder, name + ".png"))      # 1024x1024 RGB
+meta = json.load(open(os.path.join(folder, name + ".json")))
+
+print(image.size)            # (1024, 1024)
+print(meta["identity"])      # e.g. "00001"
+print(meta["face_bbox"])     # [x1, y1, x2, y2]
+print(len(meta["landmark"])) # 68
+```
+
+## Citation
+
+If you find these images useful, please cite:
+
+```bibtex
+@inproceedings{kazerouni2026face2scene,
+  title={Face2Scene: Using Facial Degradation as an Oracle for Diffusion-Based Scene Restoration},
+  author={Kazerouni, Amirhossein and Suin, Maitreya and Aumentado-Armstrong, Tristan and Honari, Sina and Walia, Amanpreet and Mohomed, Iqbal and Derpanis, Konstantinos G and Taati, Babak and Levinshtein, Alex},
+  booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
+  pages={8428--8438},
+  year={2026}
+}
+```
+
+## License
+
+The license for this dataset has not yet been finalized. Until a license is specified here, please contact the authors before redistribution or commercial use.