Search

Most existing hashing methods usually focus on constructing hash function only, rather than learning discrete hash codes directly. Therefore the learned hash function in this way may result in the hash function which can-not achieve ideal discrete hash codes. To make the learned hash function for achieving ideal approximated discrete hash codes, in this paper, we proposed a novel supervised discrete discriminant hashing learning method, which can learn discrete hashing codes and hashing function simultaneously. To make the learned discrete hash codes to be optimal for classification, the learned hashing framework aims to learn a robust similarity metric so as to maximize the similarity of the same class discrete hash codes and minimize the similarity of the different class discrete hash codes simultaneously. The discriminant information of the training data can thus be incorporated into the learning framework. Meanwhile, the hash functions are constructed to fit the directly learned binary hash codes. Experimental results clearly demonstrate that the proposed method achieves leading performance compared with the state-of-the-art semi-supervised classification methods. [ABSTRACT FROM AUTHOR]

Recently, sparse representation-based classification (SRC) has received much attention for its robustness in pattern recognition. Because SRC deals with high-dimensional data, huge computational resources are required to compute sparse representations of query samples, which renders SRC solutions of high-dimensional problems infeasible. To overcome this problem, an integrated optimisation algorithm is proposed to implement feature extraction, dictionary learning and classification simultaneously. First, to obtain sparse representation coefficients, a sparsity preserving embedding map is learnt to reduce the dimensionality of the data. Second, an optimal dictionary is adaptively obtained from the training data to reduce trivial information. Third, the training samples are reclassified using sparse representation coefficients. Furthermore, the integrated learning algorithm is extended to unsupervised learning. Experimental results clearly demonstrate that the proposed method achieves better performance than several popular feature extraction and classification methods. [ABSTRACT FROM AUTHOR]