About This Project
Developed a custom YOLOv5 object detection model for real-time detection of cars and sidelights, achieving 95% mean Average Precision (mAP). The model was integrated into the CarSight prototype for automated driving test evaluation, significantly improving the accuracy and efficiency of driving assessments.
Key Features
- Custom trained YOLOv5 model with 95% mAP
- Real-time video processing capability
- Integration with CarSight automated testing system
- Robust detection in various lighting conditions
