Infrared Object Recognition Based on Monogenic Features and Multiple Kernel Learning

Abstract

Infrared object recognition is an important branch in the field of image processing and computer vision. This paper proposes a novel infrared object recognition method based on monogenic features and multiple kernel learning. Specifically, the proposed features are mainly derived from the ideas of the monogenic signal. The applicability of the monogenic signal within the field of infrared object recognition is demonstrated by its capability of capturing both the spectral information and spatial localization with compact support. Second, to reduce the dimensionalities of the monogenic features, the principal component analysis is applied. Third, the reduced monogenic features are adaptively fused in the multiple kernel learning framework. At last, a multiple kernel learning support vector machine classifier is designed for recognizing the infrared objects. The experimental results show that the proposed method leads to good infrared object recognition performance.