13 July 2023 in IEEE Transactions on Instrumentation and Measurement
Deep learning has become a powerful tool to automatically classify medical hyperspectral images (MedHSIs) for the diagnosis of various tumors such as cancer. These deep learning based classification approaches consist of both feature extraction and disease prediction, which are independent of each other. Therefore, the extracted features may be incompatible with the used classifier for prediction. To remedy such deficiency, in this work, we propose a novel deep method, i.e ., deep margin cosine autoencoder (DMCA), for the MedHSI classification, which provides solid support to diagnose tumors. To be specific, first , we consider deep autoencoder (DAE) network as the basic framework for the feature extraction of the MedHSI, and the soft-max classifier is introduced as the output layer to predict the results. Second , to report the features well compatible with the classifier, the importance of the soft-max is added as a constraint into the DAE network. Further, a cosine margin is introduced to enhance the discrimination of different feature clusters. In addition, we also design an optimization scheme to the solutions of the DMCA model. Third , a two-stage training strategy is presented to train the built DMCA network. After completing the training, unknown tissues ( i.e ., unlabeled samples) can be specified to determine whether they are tumors or not. Sufficient experimental results have been provided to validate that our DMCA method achieves better classification performance and higher tumor diagnosis accuracy compared with some advanced approaches.
14 March 2024 in IEEE Transactions on Industrial Informatics
In this presented work, a new tracking control strategy is designed for a kind of nonholonomic system under external disturbances. First, the whole system is divided into two design stages by using the relay switching technique. Then, using state transformations, the two primitive subsystems are converted into unconstrained systems for which tracking controllers are designed based on dynamic surface control. Different from most existing finite-time control strategies and fixed-time control strategies, our controllers are able to control the tracking error within a specified accuracy within a specified time, starting from anywhere, without changing the control structure/parameters. Meanwhile, all closed-loop signals keep bounded in the whole control process. Extensive computer simulation and physical experiments have been given to validate good performance of the designed controller.
14 December 2023 in IEEE Transactions on Instrumentation and Measurement
The hyperspectral image (HSI) classification has achieved the great successes in many practical applications. In this paper, we propose a novel structure-wise ordinal measure based learning method (SOML) for the HSI classification. Specifically, the ordinal measure (OM) tool is used to construct the feature representation model, which yields the ordinal features reflecting the spectral direction information of the HSI. Additionally, the selected support vector machine (SVM) classifier is integrated into such feature representation model, which aims to guarantee the consistency between ordinal features and the SVM classifier. Further, the ϵ-dragging technique and low-rank representation are introduced for capturing the structure information of the HSI, which encourage the output ordinal features to be more separable and discriminative. We also design a proximal alternating optimization strategy to solve the constructed SOML model, and thus the optimal projection can be learned to map all samples into the ordinal features. The SVM is further used to classify these features to predict the results. Different from traditional classification methods, our proposed SOML method fully considers the information utilization of the HSI, relationship with the selected SVM classifier, and distribution of features in the stage of feature extraction, which further improve the discrimination and separability. Extensive experiments have been given to validate that the proposed SOML method achieves more competitive classification performance compared with several state-of-the-art algorithms.
1 January 2026 in IEEE Transactions on Geoscience and Remote Sensing
3 April 2026 in IEEE Signal Processing Letters
High resolution hyperspectral fusion imaging (HRHFI) is an important topic in the remote sensing related tasks, which integrates advantages of hyperspectral images (HSIs) and multipectral images (MSIs) to generate the high resolution HSIs (HR-HSIs). Traditional methods fail to maintain multi-scale change information and directional information of HSIs and MSIs during the HRHFI process, limiting the performance. To solve this challenge, a novel hierarchical gradient preserved network (HGPN) is proposed for HRHFI. Concretely, we begin by establishing the foundational observation model, which serves to make an initial estimate of HR-HSI so that the basic observation loss is thus constructed. Subsequently, we integrate both multi-scale spectral gradient preservation loss and multi-scale spatial gradient preservation loss into the basic observation loss, which results in the overall loss function of the proposed HGPN method. Further, an optimization strategy based on alternating directions is devised to solve the overall loss and derive the corresponding solutions, which finalizes the imaging fusion mapping. Different from traditional approaches, the proposed HGPN method can achieve hierarchical fusion, which retains multi-scale gradient information of input HSI/MSI for HRHFI. Extensive experiments have been implemented to verify that the proposed HGPN method achieves more competitive performance than several state-of-the-art approaches in the HRHFI task.