Docotoral Thesis Defense: Kha Gia Quach

Speaker: Kha Gia Quach

Supervisors: Drs. T. D. Bui, K. Luu

Examining Committee: Drs. A. Ben Hamza, M. Cheriet, E. Doedel, B. Jaumard, J. Y. Yu (Chair)

Title: Overcomplete Dictionary Versus Deep Learning Approaches to Image and Video Analysis

Date: Wednesday, November 8, 2017

Time: 14:00

Place: EV 3.309

ABSTRACT

Extracting useful information while ignoring others (e.g. noise, occlusion, lighting) is an essential and challenging data analyzing step for many computer vision tasks. Data analyzing of those tasks can be formulated as a form of matrix decomposition or factorization to separate useful and/or fill in missing information based on sparsity and/or low-rankness of the data. There has been an increasing number of non-convex approaches including conventional matrix norm optimizing and emerging deep learning models. However, it is hard to optimize the ideal l0-norm or learn the deep models directly and efficiently. Motivated from this challenging process, this thesis proposes two sets of approaches: conventional and deep learning based.

For conventional approaches, this thesis proposes a novel online non-convex lp-norm based Robust PCA (OLP-RPCA) approach for matrix decomposition, where 0 < p < 1. OLP-RPCA which can achieve real-time performance on large-scale data without parallelizing or implementing on a graphics processing unit is developed from the offline version LP-RPCA. A robust face recognition framework is also developed from Robust PCA and sparse coding approaches. In addition, this thesis proposes a novel Robust lp-norm Singular Value Decomposition (RP-SVD) method for analyzing two-way functional data. The proposed RP-SVD is formulated as an lp-norm based penalized loss minimization problem.

For deep learning based approaches, this thesis proposes an extension to texture modeling in the Deep Appearance Models (DAMs) by using Robust Deep Boltzmann Machines (RDBM), an alternative form of Robust Boltzmann Machines, to enhance its robustness against noise and occlusion. The extended model can cope with occlusion and extreme poses when modeling human faces in 2D image reconstruction. This thesis also introduces new fitting algorithms with occlusion awareness through the mask obtained from the RDBM reconstruction.