Image Classification by Multi-Class Boosting of Visual and Infrared Fusion with Applications to Object Pose Recognition

This paper proposes a novel method for multiview object pose classification through sequential learning and sensor fusion. The basic idea is to use images observed in visual and infrared (IR) bands, with the same sampling weight under a multi-class boosting framework. The main contribution of this paper is a multi-class AdaBoost classification framework where visual and infrared information interactively complement each other. This is achieved by learning hypothesis for visual and infrared bands independently and then fusing the optimized hypothesis subensembles. Experiments are conducted on several image datasets including a set of visual and thermal IR images containing 4844 face images in 5 different poses. Results have shown significant increase in classification rate as compared with an existing multi-class AdaBoost algorithm SAMME trained on visual or infrared images alone, as well as a simple baseline classification-fusion algorithm.