https://github.com/fengliwu/pedestrian_detection
Pedestrian detection with improved YOLOv5s model
https://github.com/fengliwu/pedestrian_detection
computer-vision object-detection yolo
Last synced: over 1 year ago
JSON representation
Pedestrian detection with improved YOLOv5s model
- Host: GitHub
- URL: https://github.com/fengliwu/pedestrian_detection
- Owner: FengliWu
- Created: 2023-12-07T17:48:55.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2023-12-28T04:37:43.000Z (over 2 years ago)
- Last Synced: 2025-01-31T19:39:03.395Z (over 1 year ago)
- Topics: computer-vision, object-detection, yolo
- Language: Python
- Homepage:
- Size: 799 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Pedestrian_Detection
## Description
An occluded pedestrian detection project using improved YOLOv5s model
## Key Features
- Improved YOLOv5s model to enhance occluded object detection capabilities by embedding SE attention mechanisms and replacing SPP with SPP-Fast modules in backbone and adding specialized detection head for small objects
- Constructed and trained improved model in PyTorch based on dataset of 2,000+ images with occluded pedestrians
## Tech Stack
- Data Processing: Python - Pandas, Numpy
- Modeling: Python - PyTorch
## Project Structure
```
Pedestrian Detection
├── data/ # trajectory dataset
├── models/
├── runs/
├── utils/
├── venv/
├── src/
| ├── detect.py # detection
| ├── export.py # export model to other formats
| ├── train.py # train
| └── val.py # validate
├── requirements.txt
├── README.md
└── LICENSE
```
Feel free to open issues and pull requests to contribute!