Human detection in video streams using a pre-trained YOLOv3 in Pytorch

A pre-trained YOLOv3 is used to detect humans(/presence of humans) in a video stream. This code can be adapted for detecting one or multiple objects in a video stream. After that, The video can be trimmed to a shorter one by avoiding all the frames that do not contain the specified objects and the audio is restituted and mixed to the resulted video.

Installation and use

Clone the repository: git clone https://github.com/hocinilotfi/Human-detection-in-video-streams-using-a-pre-trained-YOLOv3-in-Pytorch.git

Install the required packages: pip install requirements.txt

Using the script

python detect_humans.py -i inputfile -o outputfile

Example: python detect_humans.py -i input_video.mp4 -o output_video.mp4

References:

1. YOLOv3: https://pjreddie.com/darknet/yolo/
2. Erik Lindernoren's YOLO implementation: https://github.com/eriklindernoren/PyTorch-YOLOv3
3. YOLO paper: https://pjreddie.com/media/files/papers/YOLOv3.pdf