Prerequisites

1. Download Required Files

• Download YOLOv8 model put it in the resource folder:

  • YOLOv8n
  • YOLOv8s
  • YOLOv8m

• Download a test video

sudo snap install yt-dlp
yt-dlp https://youtu.be/2SATljljXCY?si=y8Lbldme2RhQADCP -f "bestvideo[ext=mp4]+bestaudio[ext=m4a]/mp4" --download-sections "*00:00:00-00:00:17" -o ./resource/sample.mp4

2. Install Required Libraries

Run the following command to install the necessary libraries:

pip install -r requirements.txt

3. Run

1. Inference using yolov8 only.

python yolov8.py

2. Inference using yolov8 + sort.

python yolov8_sort.py

4. Result

You can see the inference results during execution or check them in the result folder.

My YOLO+SORT implementation shows promising object tracking, maintaining IDs across frames. However, ID switches still occur during occlusions.

• YOLOv8 Result

• YOLOv8 + SORT Result

I plan to implement Deep SORT in future updates to improve tracking stability, especially in challenging scenarios.