This application recognises the face of a person, provided we have already trained the application with the persons face. The application can detect, track and recognise multiple faces in the video feed. We can also train this application "on the fly" to detect, recognise and track new faces.
The application first finds the face, then tracks it and then recognises the face. Even if a face is not detected in one of the frames, tracking algorithm makes sure that the face is being tracked correctly.
This project is developed on the Snowflake platform in combination with OpenCV.
Face detection, tracking and recognition with Snowflake from multitouch software on Vimeo.
Using the general model of a face to improve face recognition and expression recognition.
Let us know what you think in the comments section below.
Please note that this is a project and what you see in the video is not actually part of the Snowflake product range till date.