This project uses the YOLOv5 model for object detection on a webcam feed and overlays a trailer video on the detected objects. The trailer video is paused when certain objects are detected, and the opacity of the trailer is reduced when a person is detected. Additionally, an icon is displayed on the top right of the screen when the trailer is paused.

• Object detection using YOLOv5
• Overlay trailer video on detected objects
• Pause trailer video when certain objects are detected
• Reduce trailer opacity when a person is detected
• Display an icon when the trailer is paused

• Python 3.x
• OpenCV
• PyTorch
• YOLOv5 model

1. Clone the repository:

   git clone git@github.com:char-projects/spatial-computing.git spatialComputing
   cd ./spatialComputing/eyeContact

2. Install the required dependencies:

   pip install -r requirements.txt

1. Ensure you have a trailer.mp4 file in the project directory. You can change the source to any other video file if needed.
2. Run the script:

   python3 eyeContact/test.py

• To change the camera source, modify the device index in the code:

  cap = cv2.VideoCapture(0)  # Change to the correct device index for your camera

• To change the video file, update the path in the code:

  video_path = os.path.abspath('trailer.mp4')  # Replace with your video file path

• To change the icon image, update the path in the code:

  icon_path = os.path.abspath('icon.png')  # Replace with your icon image path

• To modify the object classes that pause the trailer or reduce opacity, update the lists in the code:

  stop_objects = ['car', 'bicycle']  # Add the object classes you want to stop the video
  reduce_opacity_objects = ['person']  # Add the object classes you want to reduce the opacity