Add model card with description, usage, and benchmarks

README.md

@@ -0,0 +1,81 @@
+---
+language: en
+license: mit
+tags:
+- document-forgery-detection
+- tampering-detection
+- image-manipulation
+- vision-transformer
+- lora
+pipeline_tag: image-segmentation
+datasets:
+- DocTamperV1
+metrics:
+- f1
+---
+
+# SEED — Simple & Efficient Expert Detector
+
+**SEED** (Simple & Efficient Expert Detector) is a lightweight vision transformer model for document forgery detection. It localizes tampered regions in document images and classifies images as real or forged.
+
+## Architecture
+
+| Component | Detail |
+|-----------|--------|
+| Backbone | [DINOv3 ViT-L/16](https://huggingface.co/facebook/dinov3-vitl16-pretrain-lvd1689m) |
+| Finetuning | LoRA (rank=1, attention + MLP) |
+| Queries | 1 mask query |
+| Decoder blocks | 4 |
+| Input size | 512 × 512 |
+| Parameters | ~304M (only ~1M trainable with LoRA) |
+
+## Usage
+
+```python
+from model.hf_wrapper import EoMTForTamperingDetection
+
+model = EoMTForTamperingDetection.from_pretrained("Jason37437/SEED")
+model.eval()
+
+# The model outputs:
+#   - mask_logits: per-query segmentation masks
+#   - class_logits: per-query foreground/background scores  
+#   - image_logits: image-level real vs forged classification
+```
+
+## Performance
+
+Evaluated on 1,000 samples per dataset at 512×512:
+
+### Detection (image-level F1)
+| Dataset | F1 | Accuracy | Precision | Recall |
+|---------|-----|----------|-----------|--------|
+| T-SROIE | 0.88 | 0.88 | 0.90 | 0.86 |
+| Tampered-IC13 | 0.91 | 0.95 | 0.98 | 0.86 |
+| RealTextManipulation | 0.36 | 0.63 | 0.49 | 0.28 |
+| OSTF | 0.57 | 0.92 | 0.56 | 0.58 |
+
+### Localization (pixel-level F1)
+| Dataset | F1 |
+|---------|-----|
+| T-SROIE | 0.39 |
+| Tampered-IC13 | 0.26 |
+| RealTextManipulation | 0.04 |
+| OSTF | 0.05 |
+
+> **Note:** This is a lightweight LoRA rank-1 model trained for 5,000 steps. Higher localization performance can be achieved with larger LoRA ranks and longer training.
+
+## Citation
+
+```bibtex
+@software{seed2025,
+  title={SEED: Simple & Efficient Expert Detector for Document Forgery},
+  author={Mobile Perception Systems Lab, TU/e},
+  year={2025},
+  url={https://huggingface.co/Jason37437/SEED}
+}
+```
+
+## License
+
+MIT License.