Please use this identifier to cite or link to this item: http://hdl.handle.net/11023/1823
Title: Wavelet-based Recognition of Facial Expressions and Faces
Author: Poursaberi, Ahmad
Advisor: Yanushkevich, Svetlana
Gavrilova, Marina
Keywords: Engineering--Electronics and Electrical
Issue Date: 29-Sep-2014
Abstract: The proposed research contributes to the development of biometric systems and its applications. The facial biometric, as a part of the underlying technology, is a focus of this thesis. Specifically, it is aimed at developing the new concepts of biometrics systems, such as systems with situational awareness, interviewing, and human-machine interaction support system that analyze human behavioral biometrics. The systems make use of facial expression data, besides the traditional visual-spectrum data for face recognition. A new technique for facial expression recognition (FER), in particular, is targeted in this thesis as one of the important components in human behavior analysis. Combined with face recognition (FR), it is a basis for a multi-biometric decision support tool. Most of the available approaches for the recognition of faces and expressions only consider either the expression-invariant face recognition or the facial expression recognition regardless of identity. This research proposes a facial biometric analysis using a new facial feature detection based on a modified Frangi filter. The joint FR/FER uses new feature extraction technique called Gauss-Laguerre wavelets, which provides rich and unique features suitable for expressions classification and identity recognition. In addition, this study considers facial biometric in infrared spectrum, which provides additional information (such as face region temperature) which is required, in particular, at a border check point. Moreover, recognition of expressions from video is investigated to show the extendibility of the proposed technique for video processing as well.
URI: http://hdl.handle.net/11023/1823
Appears in Collections:Electronic Theses

Files in This Item:
File Description SizeFormat 
Ucalgary_2014_Poursaberi_Ahmad.pdf5.29 MBAdobe PDFView/Open


Items in The Vault are protected by copyright, with all rights reserved, unless otherwise indicated.