Semi-supervised Bi-dictionary Learning Using Smooth Representation-Based Label Propagation

摘要

Due to heavy clutters and occlusions of complex background, natural images contain complex features in data structure which often cause errors in image classification. In this paper, we propose semi-supervised bi-dictionary learning for image classification with smooth representation-based label propagation (SRLP) which extends reconstruction-based classification in a probabilistic manner. First, we jointly learn a discriminative dictionary in the feature space and its corresponding soft label in the label space. Then, we utilize the learnt bi-dictionary in image classification based on SRLP. Experimental results demonstrate that the proposed SRLP is capable of learning the discriminative bi-dictionary for image classification and outperforms the-state-of-the-art reconstruction-based classification methods.