Alumni

Facial recognition and the noisy communication channel

View Full Record
Description: The Coding Theory concept of the noisy communication channel is examined as a useful model of facial recognition. The Bayesian facial recognition technique proposed by Moghaddam, Jebara, and Pentland [Mogh00] is examined in terms of the noisy communication channel metaphor, resulting in the proposition of an algorithm that explicitly applies the Bayesian facial recognition techniques to facial recognition using the noisy communication channel model. The proposed algorithm is tested on a database containing common image variations known to make face recognition particularly difficult (such as the presence of sunglasses or a large scarf). This initial testing indicates that the proposed algorithm yields improvement on the Bayesian technique. Finally, the suitability of the noisy communication channel metaphor is found to be a promising avenue for future research.* *This dissertation is a compound document (contains both a paper copy and a CD as part of the dissertation). The CD requires the following system requirements: Adobe Acrobat.
Language: English
Format: Degree Work