‘Hands-free Interface’ – A Fast and Accurate Tracking Procedure for Real Time Human Computer Interaction

This paper introduces a method to track features using image recognition to be used as in input device.

The feature chosen for tracking is the nostril. This is due to its distinct shape, color contrast, and size.

Template matching (correlation based) is used to track the feature. The normalized sum of squared (NSSD) is used to compare the gray level intensities in the new and old image in order to find the desired feature.

This paper also addresses two problems with template matching; image distortion (due to rotation or scaling) and when the  tracked feature becomes blocked.

To increase success, the current image is compared to both the previous image and the initial image.  The results are scaled to favor the previous image (by .25) and the minimum is chosen to be the new location.

This limits the feature draft and image timeouts.

A Kalman filter is also used to reduce search space.

Nodding and face shaking was also recognized.

Signal Processing and Information Technology, 2004. Proceedings of the Fourth IEEE International Symposium on
Date of Conference: 18-21 Dec. 2004
Author(s): El-Afifi, L. 
Electr. & Comput. Eng. Dept., American Univ. of Beirut
Karaki, M. ;  Korban, J. ;  Alaoui, M.A.A. 
Page(s): 517 – 520
Product Type: Conference Publications

About Frank Sposaro

Frank was the initial student to start the mobile lab with Dr. Tyson. After working on the first project, iFall, he and Dr. Tyson designed the Mobile Programming course as FSU. The course is used as a training base to recruit new students into the lab. His thesis researches several medical related applications, including iFall. Frank then went on to implement the redesign of the “favorite contacts” for Android’s Ice Cream Sandwich at Google HQ in Mountain View, California. He currently acts as a tech lead in the lab getting infrastructure and project management tools setup. He has special focus on native Android coding and UI design.