Computer Vision Based Application Control

Description:

This project includes different types of computer-application control without keyboard and mouse using computer-vision algorithms. A webcam is required with the computer for this purpose. We have used Logitech c310 webcam.

It includes:

  • Motion-tracking based gaming. This enables the user to play games without any keyboard, mouse and game-controllers. It requires only one webcam to detect different motions and generate actions according to it.
  • Face-tracking based image viewing. This enables the user to enjoy panoramic pictures without using tedious mouse-movements. A webcam is required to track the face and the view transforms itself according to facial position.
  • Hand-movement based slide-show. No spacebar press or mouse-click is required for presenting a slide. Only a webcam is required. It tracks the hand-movement and changes the slide accordingly.

For motion detection, optical flow is used.

For face-tracking, at first the face is detected using cascade object detector. Cascade object detector uses the Viola-Jones detection algorithm and a trained classification model for detection. Then facial feature is identified to track. Here, facial tone is used as the facial feature. Then face is tracked using Histogram based tracking.

The code is shared in my github page in this link: https://github.com/mdshafayat/motion_track_game