GateDetection is a collection of computer vision-based gate detection approaches implemented in Python and Jupyter Notebooks. This project includes multiple detection methods such as:
- YOLO-based detection
- RCNN-based detection
- Color matching techniques
- Dataset : https://www.kaggle.com/datasets/mobilal14/sauvc-dataset
- models : https://huggingface.co/Bilal1410/GateDetection-model
- Madras Dataset : https://app.roboflow.com/main-j13ii/gatedetection-ujxkq/3
| Model | mAP@0.5 | mAP@0.5:0.95 | Precision | Recall |
|---|---|---|---|---|
| yolo_v11n | 0.9456 | 0.7671 | 0.9679 | 0.9161 |
| yolo_v5n | 0.9386 | 0.7652 | 0.9644 | 0.9051 |
| yolo_v8n | 0.9319 | 0.7644 | 0.9732 | 0.8875 |
| yolo_v8m | 0.9683 | 0.7897 | 0.9670 | 0.9454 |
-
These results are obtained on the GateDetection Dataset (https://www.kaggle.com/datasets/mobilal14/sauvc-dataset) using Tesla P100-PCIE-16GB GPU.
-
Model Comparison Notebook : model-comparison.ipynb