Class mean‐weighted discriminative collaborative representation for classification

Abstract

Representation-based classification (RBC) has been attracting a great deal of attention in pattern recognition. As a typical extension to RBC, collaborative representation-based classification (CRC) has demonstrated its superior performance in various image classification tasks. Ideally, we expect that the learned class-specific representations for a testing sample are discriminative, and the representation computed for the true class dominates the final representation of the testing sample. Most existing CRC-based methods can learn pattern discrimination, but cannot differentiate the contribution of class-specific representations to the classification of each testing sample. It is challenging for a representation-based classifier to retain both properties. To address this challenge and further improve CRC's classification performance, we propose a novel CRC-based method, class mean-weighted discriminative collaborative representation-based classifier (CMW-DCRC). Its objective function penalises the standard l 2 -norm residuals with two discriminative regularisation terms. A decorrelating term makes the class-specific representations more discriminative, and a newly designed class meanweighted term that promotes the training samples from individual classes to competitively reconstruct the testing sample while boosting the contribution of the true class. To further enhance the robustness of CRC, we extend CMW-DCRC by replacing the l 2 -norm coding residual with a l 1 -norm coding residual, and solve the optimisation problem with an iteratively reweighted least square algorithm. Extensive experimental results on nine image data sets have shown that our methods outperform the state-of-the-art RBC-based methods.

An intelligent system for craniomaxillofacial defecting reconstruction

“…In recent years, with the successful application of deep learning in the field of medical imaging, [21][22][23][24][25][26] more and more people have tried to apply deep learning to virtual reconstruction, and many results have been achieved. [27][28][29] Morais et al 27 proposed a denoising auto-encoder.…”

Book Classification and Recommendation for University Libraries using Grey Correlation and Bayesian Probability

“…Classifiers based on collaborative representation have been applied to many practical cognitive domains due to their advantages in terms of efficiency and effectiveness. A new neighbourhood prior constrained collaborative representation model was suggested by Gou et al The guidance of the neighbourhood prior in the encoding process was emphasized [16]. Hyper-spectral pictures' extensive spectral data enabled several applications with tremendous advantages.…”

Robust visual tracking via adaptive feature channel selection

“…For example, for image classification tasks, a CNN model needs to learn the features of thousands of target categories and save them in various channels in different layers. 21 However, we need to pour more attention into the features of specific targets in the visual tracking task, so the feature channels irrelevant to the current target are not necessary. Second, whether it is the filter learning process or the target positioning process, the increase of the feature dimension will cause a more significant computational burden and reduce the practicability of the algorithm when solving the DCF formulation.…”

“…The CNN models widely used in visual tracking are pretrained on the large‐scale data set ImageNet, 20 which is mainly designed for image classification tasks and is not fully applicable to visual tracking tasks. For example, for image classification tasks, a CNN model needs to learn the features of thousands of target categories and save them in various channels in different layers 21 . However, we need to pour more attention into the features of specific targets in the visual tracking task, so the feature channels irrelevant to the current target are not necessary.…”