Discriminative models for robust image classification

A variety of real-world tasks involve the classification of images into pre-determined categories. Designing image classification algorithms that exhibit robustness to acquisition noise and image distortions, particularly when the available training data are insufficient to learn accurate models, is a significant challenge. This dissertation explores the development of discriminative models for robust image classification that exploit underlying signal structure, via probabilistic graphical models and sparse signal representations. Probabilistic graphical models are widely used in many applications to approximate high-dimensional data in a reduced complexity set-up. Learning graphical structures to approximate probability distributions is an area of active research. Recent work has focused on learning graphs in a discriminative manner with the goal of minimizing classification error. In the first part of the dissertation, we develop a discriminative learning framework that exploits the complementary yet correlated information offered by multiple representations (or projections) of a given signal/image. Specifically, we propose a discriminative tree-based scheme for feature fusion by explicitly learning the conditional correlations among such multiple projections in an iterative manner. Experiments reveal the robustness of the resulting graphical model classifier to training insufficiency. The next part of this dissertation leverages the discriminative power of sparse signal representations. The value of parsimony in signal representation has been recognized for a long time, most recently in the emergence of compressive sensing. A recent significant contribution to image classification has incorporated the analytical underpinnings of compressive sensing for classification tasks via class-specific dictionaries. In continuation of our theme of exploiting information from multiple signal representations, we propose a discriminative sparsity model for image classification applicable to a general multi-sensor fusion scenario. As a specific instance, we develop a color image classification framework that combines the complementary merits of the the red, green and blue channels of color images. Here signal structure manifests itself in the form of block-sparse coefficient matrices, leading to the formulation and solution of new optimization problems. As a logical consummation of these ideas, we explore the possibility of learning discriminative graphical models on sparse signal representations. Our efforts are inspired by