Update README.md

README.md

@@ -1,10 +1,220 @@
----
-title: README
-emoji: 🌍
-colorFrom: purple
-colorTo: pink
-sdk: static
-pinned: false
+---
+language:
+- en
+license:
+- cc-by-4.0
+tags:
+- semantic-segmentation
+- scene-understanding
+- 3d-point-clouds
+- multimodal
+- drone-imagery
+pretty_name: NEST3D - Sociable Weaver Nest 3D Dataset
+title: 'NEST3D: A High-Resolution Multimodal Dataset of Sociable Weaver Tree Nests '
+---
+
+# NEST3D: A High-Resolution Multimodal Dataset of Sociable Weaver Tree Nests
+
+![NEST3D Workflow](./workflow.png)
+
+## Dataset Description
+
+NEST3D is a multimodal dataset of 104 sociable weaver nests, combining drone-based RGB and multispectral imagery with a semantically annotated 3D RGB point cloud. It captures trees hosting these nests through drone-based remote sensing, providing rich spatial and spectral information to benchmark and advance scene-level semantic segmentation methods for computer vision and ecological monitoring applications.
+
+### Key Characteristics
+
+- **Modality**: Multimodal (RGB imagery, multispectral bands, 3D point clouds)
+- **Task**: Scene-level semantic segmentation
+- **Scale**: Multiple tree-nest scenes with consistent spatial and spectral coverage
+- **Annotation**: Point-level semantic labels for 3D point clouds
+- **Data Source**: Drone-based RGB and multispectral imagery
+- **Application Domain**: Ecological monitoring, wildlife management, 3d semantic segmenation, 3d reconstruction.
+
+## Dataset Organization
+
+The dataset is organized into modality-specific directories to support flexible access and reuse:
+
+### Directory Structure
+
+```
+NEST3D/
+├── train/
+│   ├── sample_001/
+│   │   ├── RGB/                    # RGB drone images
+│   │   │   ├── sample001_RGB_001.JPG
+│   │   │   └── ...
+│   │   ├── MS/                     # Multispectral imagery
+│   │   │   ├── Green/
+│   │   │   │   ├── sample001_G_001.TIF
+│   │   │   │   └── ...
+│   │   │   ├── Red/
+│   │   │   │   ├── sample001_R_001.TIF
+│   │   │   │   └── ...
+│   │   │   ├── Red_Edge/
+│   │   │   │   ├── sample001_RE_001.TIF
+│   │   │   │   └── ...
+│   │   │   └── NIR/
+│   │   │       ├── sample001_NIR_001.TIF
+│   │   │       └── ...
+│   │   └── sample001.npy            # 3D point cloud with labels
+│   └── sample_002/
+│       └── ...
+│
+└── test/
+    ├── sample_084/
+    │   ├── RGB/
+    │   ├── MS/
+    │   └── sample084.npy
+    └── ...
+```
+
+### Data Modalities
+
+#### 1. **RGB Imagery**
+- Raw drone images from aerial acquisition
+- Format: JPEG
+- Organized by data split and scene identifier
+- Example path: `train/sample_001/RGB/sample001_RGB_119.JPG`
+
+#### 2. **Multispectral Imagery**
+- Four spectral bands from the same acquisitions as RGB
+- Organized into four band-specific folders:
+  - **Green (G)**: Green channel imagery
+  - **Red (R)**: Red channel imagery
+  - **Red Edge (RE)**: Red Edge channel for vegetation analysis
+  - **NIR**: Near-Infrared channel for vegetation health assessment
+- Format: GeoTIFF (.TIF)
+- Example paths:
+  - `train/sample_001/MS/Green/sample001_G_119.TIF`
+  - `train/sample_001/MS/Red/sample001_R_119.TIF`
+  - `train/sample_001/MS/Red_Edge/sample001_RE_119.TIF`
+  - `train/sample_001/MS/NIR/sample001_NIR_119.TIF`
+
+#### 3. **3D Point Clouds**
+- One NumPy file per scene containing the complete 3D reconstruction
+- Format: `.npy` (NumPy binary format)
+- Per-point attributes: `[x, y, z, r, g, b, label]`
+  - **x, y, z**: 3D spatial coordinates (meters)
+  - **r, g, b**: RGB color values (0-255)
+  - **label**: Semantic class label (integer)
+- Example path: `train/sample_001/sample001.npy`
+
+## Data Splits
+
+The dataset is divided into fixed training and test sets:
+
+- **Training Set**: Used for model training and development
+- **Test Set**: Reserved for model evaluation and benchmarking
+
+Each split contains a consistent collection of scenes to ensure reliable evaluation.
+
+## Usage
+
+### Loading 3D Point Clouds
+
+```python
+import numpy as np
+
+# Load point cloud with semantic labels
+point_cloud = np.load('train/sample_001/sample001.npy')
+
+# Extract coordinates
+xyz = point_cloud[:, :3]
+
+# Extract colors
+rgb = point_cloud[:, 3:6]
+
+# Extract semantic labels
+labels = point_cloud[:, 6]
+```
+
+### Loading Multispectral Imagery
+
+```python
+from PIL import Image
+import numpy as np
+
+# Load a single band
+green_band = np.array(Image.open('train/sample_001/MS/Green/sample001_G_001.TIF'))
+
+# Load all four bands for a given image
+green = np.array(Image.open('train/sample_001/MS/Green/sample001_G_001.TIF'))
+red = np.array(Image.open('train/sample_001/MS/Red/sample001_R_001.TIF'))
+red_edge = np.array(Image.open('train/sample_001/MS/Red_Edge/sample001_RE_001.TIF'))
+nir = np.array(Image.open('train/sample_001/MS/NIR/sample001_NIR_001.TIF'))
+
+# Stack into multiband image
+multispectral = np.stack([green, red, red_edge, nir], axis=-1)
+```
+
+### Using with Hugging Face Datasets Library
+
+```python
+from datasets import load_dataset
+
+# Load the dataset
+dataset = load_dataset('NEST3D/dataset')
+```
+
+## Downloading the Dataset
+
+### Option 1: Using Hugging Face Hub
+```bash
+pip install huggingface_hub
+
+huggingface-cli download NEST3D/dataset --repo-type dataset --local-dir ./NEST3D
+```
+
+### Option 2: Direct Download
+Visit [https://huggingface.co/datasets/NEST3D/dataset](https://huggingface.co/datasets/NEST3D/dataset) and download files directly.
+
+## Citation
+
+If you use this dataset in your research, please cite:
+
+```bibtex
+@dataset{NEST3D,
+  title={NEST3D: A High-Resolution Multimodal Dataset of Sociable Weaver Tree Nests},
+  organization={NEST3D},
+  url={https://huggingface.co/datasets/NEST3D},
+  year={2026}
+}
+```
+
+## License
+
+This dataset is released under the Creative Commons Attribution 4.0 International (CC-BY-4.0) License. You are free to use, distribute, and adapt the dataset as long as you provide appropriate attribution.
+
+See [LICENSE](LICENSE) for details.
+
+## Dataset Information
+
+- **Total Size**: 
+- **Number of Scenes**: 104 samples
+- **Modalities**: RGB, Multispectral (4 bands), 3D Point Clouds
+- **Image Format**: JPEG (RGB), GeoTIFF (Multispectral)
+- **Point Cloud Format**: NumPy arrays
+- **Annotation Type**: Per-point semantic labels
+
+## Acknowledgments
+
+This work was funded by:
+
+- **European Union's Horizon Europe** research and innovation programme through the Marie Skłodowska-Curie project **"WildDrone – Autonomous Drones for Nature Conservation"** (grant agreement no. 101071224)
+- **EPSRC-funded** "Autonomous Drones for Nature Conservation Missions" grant (EP/X029077/1)
+- **Swiss State Secretariat for Education, Research and Innovation (SERI)** under contract number 22.00280
+
+We extend our gratitude to our collaborators and field partners in Namibia for their invaluable support during data collection.
+
+## Contact & Support
+
+For questions, issues, or contributions, please visit the [dataset discussion forum](https://huggingface.co/datasets/NEST3D/dataset/discussions).
+
+## Disclaimer
+
+The NEST3D dataset is provided as-is for research and development purposes. Users are responsible for ensuring their use complies with all applicable laws and regulations, particularly regarding ecological monitoring and wildlife protection.
+
 ---
 
-Edit this `README.md` markdown file to author your organization card.
+**Last Updated**: February 2026  
+**Dataset Version**: 1.0