Date of Graduation


Document Type


Degree Type



Statler College of Engineering and Mineral Resources


Lane Department of Computer Science and Electrical Engineering

Committee Chair

Vinod Kulathumani


Recent terror attacks, intrusion attempts and criminal activities have necessitated a transition to modern biometric systems that are capable of identifying suspects in real time. But real-time biometrics is challenging given the computationally intensive nature of video processing and the potential occlusions and variations in pose of a subject in an unconstrained environment. The objective of this dissertation is to utilize the robustness and parallel computational abilities of a distributed camera network for fast and robust face recognition.;In order to support face recognition using a camera network, a collaborative middle-ware service is designed that enables the rapid extraction of multi-view face images of multiple subjects moving through a region. This service exploits the epipolar geometry between cameras to speed up multi view face detection rates. By quickly detecting face images within the network, labeling the pose of each face image, filtering them based on their suitability of recognition and transmitting only the resultant images to a base station for recognition, both the required network bandwidth and centralized processing overhead are reduced. The performance of the face image acquisition system is evaluated using an embedded camera network that is deployed in indoor environments that mimic walkways in public places. The relevance of the acquired images for recognition is evaluated by using a commercial software for matching acquired probe images. The experimental results demonstrate significant improvement in face recognition system performance over traditional systems as well as increase in multi-view face detection rate over purely image processing based approaches.