Edge_Vision

Introduction

We are actively conducting research focused on the design and optimization of lightweight AI models capable of running efficiently on edge devices. Our primary goal is to develop high-performance AI solutions for real-time perception in autonomous driving scenarios. Key research activities include leveraging edge devices such as the Jetson AGX Orin to implement and optimize these models. By utilizing state-of-the-art optimization techniques such as Neural Architecture Search (NAS), Pruning, and Quantization, we aim to minimize memory usage and inference time while maintaining stable performance. Additionally, our research explores hardware acceleration and parallel processing tools to identify and address performance bottlenecks, providing optimal solutions tailored for autonomous driving systems.

Use in a Local

YOLO-series

YOLOv5 ~ v11 for Inference at Local & Jetson board

Installation

  pip install ultralytics  

  pip install -r requirements.txt  

yolov5 github

https://github.com/ultralytics/yolov5

Inference using webcam

TBD

yolov6 github

https://github.com/meituan/YOLOv6

Inference using webcam

  python3 tools/infer.py --weights yolov6n.pt --source 0 --webcam --yaml data/coco.yaml --view-im  

yolov7 github

https://github.com/WongKinYiu/yolov7

Inference using webcam

  python3 detect.py --weights yolov7.pt --conf 0.25 --source 0 --view-img --view-im  

yolov8 github

https://github.com/autogyro/yolo-V8 https://docs.ultralytics.com/ko/models/yolov8/

Inference using webcam

  python3 infer.py --weights yolov8n.pt --conf 0.25 --source 0 --view-img --view-im  

yolov9 github

https://github.com/WongKinYiu/yolov9

Inference using webcam

  python3 detect.py --weights yolov9t.pt --conf 0.25 --source 0 --view-img  

yolov10 github

https://github.com/THU-MIG/yolov10

Inference using webcam

  python3 app.py --weights yolov10n.pt --conf 0.25 --source 0 --view-img  

and also connect to the link below

  http://127.0.0.1:7860  

yolov11 github

https://github.com/THU-MIG/yolov10

Inference using webcam

  python3 infer.py --weights yolov11n.pt --conf 0.25 --source 0 --view-img  

Use in a AGX Orin

TBD

Use in a Orin Nano

TBD

Contact

Vehicle Intelligence Perception LAB at Gachon University https://sites.google.com/view/vip-lab

• Jhonghyun An, Professor of the Gachon Univ / jhonghyun@gachon.ac.kr
• Nahyeong Kim, Master's Student of the Gachon Univ / skgud99@gachon.ac.kr
• Seong Kyu Choi, Master's Student of the Gachon Univ / seongkyu950324@gachon.ac.kr
• Young Jae Cheong, Master's Student of the Gachon Univ / bluebull777@gachon.ac.kr
• Yejun Lee, Master's Student of the Gachon Univ / dldpwns001@gachon.ac.kr
• Sungju An, Master's Student of the Gachon Univ / samsj1209@gachon.ac.kr
• Siwoo Kim, Master's Student of the Gachon Univ / ksw5830@gachon.ac.kr

---