Overview:
This project focuses on detecting protective equipment such as helmets and jackets using the YOLOv8m model. It processes 3235 images and classifies seven different protective equipment types to enhance safety in environments like construction sites. The model run 90 epochs of training with rotation-based augmentation for improved detection accuracy.
Key Features:
1. YOLOv8m for Protective Equipment Detection:
The model detects seven classes of protective gear, including helmets and jackets, ensuring high accuracy in identifying essential safety equipment.
2. Dataset of 3235 Images:
The model was trained on 3235 images with a resolution of 416×416, representing a variety of protective gear in different environments.
3. Augmentation:
The dataset was enhanced with rotation-based augmentation, helping the model improve accuracy across various angles and perspectives.
4. Performance:
The model achieved a mean average precision (mAP) of 0.902, with a precision of 0.911 and recall of 0.865, indicating reliable detection.
5. Potential Improvement: Balancing data samples across all classes could further boost model performance, as some classes may be underrepresented in the dataset.
Category:
Tags:
Links
Portfolio 
Leave a comment