|
|
--- |
|
|
license: apache-2.0 |
|
|
tags: |
|
|
- computer-vision |
|
|
- chess |
|
|
- yolo |
|
|
- object-detection |
|
|
- piece-detection |
|
|
|
|
|
model-index: |
|
|
- name: chess_piece_detection |
|
|
results: [] |
|
|
--- |
|
|
|
|
|
# chess_piece_detection |
|
|
|
|
|
Chess piece detection model trained with YOLO |
|
|
|
|
|
## Model Details |
|
|
|
|
|
- **Model Type**: YOLO Detection |
|
|
- **Task**: object-detection |
|
|
- **License**: apache-2.0 |
|
|
|
|
|
## Usage |
|
|
|
|
|
### Loading the Model |
|
|
|
|
|
```python |
|
|
from src.chess_piece_detection.model import ChessModel |
|
|
|
|
|
# Load the model |
|
|
model = ChessModel(model_path="path/to/downloaded/model.pt") |
|
|
|
|
|
# Detect chess pieces in an image |
|
|
detected_pieces = model.detect_pieces("path/to/chessboard_image.jpg") |
|
|
|
|
|
for square_num, piece_class in detected_pieces: |
|
|
piece_name = model.get_piece_name(piece_class) |
|
|
print(f"Square {square_num}: {piece_name}") |
|
|
|
|
|
# Visualize results |
|
|
model.visualize_detections("path/to/chessboard_image.jpg", show=True) |
|
|
``` |
|
|
|
|
|
### Direct YOLO Usage |
|
|
|
|
|
```python |
|
|
from ultralytics import YOLO |
|
|
|
|
|
# Load the model |
|
|
model = YOLO("path/to/downloaded/model.pt") |
|
|
|
|
|
# Run detection on chess piece images |
|
|
results = model("path/to/chessboard_image.jpg") |
|
|
|
|
|
# Get bounding boxes and classifications |
|
|
for result in results: |
|
|
boxes = result.boxes |
|
|
if boxes is not None: |
|
|
for box in boxes: |
|
|
class_id = int(box.cls[0]) |
|
|
confidence = float(box.conf[0]) |
|
|
print(f"Detected piece class {class_id} with confidence {confidence:.2f}") |
|
|
``` |
|
|
|
|
|
### Training Data Format |
|
|
|
|
|
This model expects YOLO detection format with chess piece annotations: |
|
|
|
|
|
```yaml |
|
|
# data.yaml |
|
|
train: path/to/train/images |
|
|
val: path/to/val/images |
|
|
nc: 12 # Number of chess piece classes |
|
|
names: ['white-king', 'white-queen', 'white-rook', 'white-bishop', 'white-knight', 'white-pawn', |
|
|
'black-king', 'black-queen', 'black-rook', 'black-bishop', 'black-knight', 'black-pawn'] |
|
|
``` |
|
|
|
|
|
With corresponding label files containing piece bounding boxes: |
|
|
``` |
|
|
# labels/image.txt |
|
|
class_id x_center y_center width height # normalized coordinates |
|
|
``` |
|
|
|
|
|
## Training |
|
|
|
|
|
This model was trained using the Chess Piece Detection training pipeline: |
|
|
|
|
|
```bash |
|
|
python src/chess_piece_detection/train.py \ |
|
|
--data data/chess_pieces/data.yaml \ |
|
|
--epochs 100 \ |
|
|
--batch 16 \ |
|
|
--img-size 640 |
|
|
``` |
|
|
|
|
|
## Model Performance |
|
|
|
|
|
<!-- Add performance metrics here after training --> |
|
|
|
|
|
## Citation |
|
|
|
|
|
If you use this model in your research, please cite: |
|
|
|
|
|
```bibtex |
|
|
@misc{dopaul_chess_piece_detection, |
|
|
title={chess_piece_detection}, |
|
|
author={dopaul}, |
|
|
year={2024}, |
|
|
publisher={Hugging Face}, |
|
|
url={https://huggingface.co/dopaul/chess_piece_detection} |
|
|
} |
|
|
``` |
|
|
|
