Class Incremental Learning (CIL) aims to enable models to continually learn new classes while retaining previously learned knowledge. The principal challenge in CIL is catastrophic forgetting, which prior approaches typically address by distilling knowledge from previous model. However, such way is often limited to pairwise alignment, failing to preserve the underlying global manifold structure of feature space—ultimately resulting in semantic drift over time. To capture multi-scale structural patterns in the feature space, we propose a topology-aware distillation framework that leverages persistent homology. Specifically, by enforcing topological alignment across incremental stages, our method ensures structure-consistent knowledge transfer and robust preservation of old classes. Furthermore, we still devise a dual-branch architecture with an inverse sampling and dynamic reweighting mechanism that addresses the inherent data imbalance in standard replay-based frameworks. These innovations coalesce into TaKP (Topology-aware Knowledge Preservation), a unified framework designed to enhance knowledge preservation in CIL. Extensive experiments demonstrate that TaKP achieves state-of-the-art performance on multiple benchmarks, significantly improving old-class preservation and average accuracy.

Few-Shot Class-Incremental Learning (FSCIL) aims to continuously learn new classes from a limited set of training samples without forgetting knowledge of previously learned classes. Conventional FSCIL methods typically build a robust feature extractor during the base training session with abundant training samples and subsequently freeze this extractor, only fine-tuning the classifier in subsequent incremental phases. However, current strategies primarily focus on preventing catastrophic forgetting, considering only the relationship between novel and base classes, without paying attention to the specific decision spaces of each class. To address this challenge, we propose a plug-and-play Adaptive Decision Boundary Strategy (ADBS), which is compatible with most FSCIL methods. Specifically, we assign a specific decision boundary to each class and adaptively adjust these boundaries during training to optimally refine the decision spaces for the classes in each session. Furthermore, to amplify the distinctiveness between classes, we employ a novel inter-class constraint loss that optimizes the decision boundaries and prototypes for each class. Extensive experiments on three benchmarks, namely CIFAR100, miniImageNet, and CUB200, demonstrate that incorporating our ADBS method with existing FSCIL techniques significantly improves performance, achieving overall state-of-the-art results.

Fine-grained vehicle classification, which is a key technology within intelligent transportation systems, has been gaining increasing importance with the burgeoning growing number of vehicles. Previous studies have predominantly focused on intricate and distinctive local features. However, in various tasks, it has been proven that global features are of significant importance when they can be effectively integrated with local features in a harmonious manner. So, we consider that a comprehensive consideration of both local and global features is crucial for enhancing classification decisions. Consequently, the paper designs a novel architecture for the task, which combines global and local features to improve classification performance. The architecture consists of two components: the local-feature net and the global-feature net. Specially, for the local feature, we propose an Essential Part Locator module that uses global feature-weighted attention masks to obtain local features, and a Cross-Part Feature Transformer that boosts interactions between local features. Meanwhile, our architecture processes the entire image through an encoder to capture global features and then integrates both global and local features. Experimental results on the Stanford Cars, CompCars, and BoxCars116K datasets demonstrate that the proposed approach surpasses state-of-the-art methods, achieving accuracies of 97.5%, 96.4%, and 92.1%, respectively.

Mild cognitive impairment (MCI) is the early stage of Alzheimer's disease, which is associated with abnormal brain proteins, the recognition of MCI being a challenging task. Recent studies have shown that the performance of MCI identification can be improved by combining protein features captured in Positron Emission Computed Tomography(PET). Nevertheless, there are still great challenges in extracting effective features from the vast amount of information. Most brain networks only considered the unilateral features of nodes or edges, ignored the interactions between them. In response to this problem, our study proposed to combine the quadruple Siamese network and GCN with self-attention pooling(QS-SAGCN) for MCI identification. In detail, we constructed the multiple protein features network(MPN) and higher-order MPN(MPHN) by PET images to promote the MCI identification. Furthermore, a pooling operation with self-attention mechanism was incorporated into GCN(SAGCN), which considered the node characteristics and topology in the graph network to facilitate the acquisition of robust biomarkers, simultaneously. Additionally we combined quadruple Siamese network with SAGCN as classification framework to improve the identification accuracy. Our proposed MCI identification method was evaluated on 230 subjects (including 117 MCI subjects, 113 normal control subjects) with both 18F-AV-1451 PET and 18F-AV-45 PET data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. Experimental results showed that 1) QS-SAGCN enhanced the ability of feature identification, laying the foundation for obtaining more effective biomarkers for MCI patients; 2) The MCI identification accuracy (93.5%) was obtained by combining QS-SAGCN and higher-order network, indicating that the framework had advantages in mental disorders recognition. Finally, through comparison, the accuracy of our proposed MCI recognition method was superior to some of the existing state-of-the-art methods. Overall, the MCI identification method in this study was effective and promising to assist in the diagnosis of MCI in future clinical practice.

Training noise-robust deep neural networks (DNNs) in label noise scenario is a crucial task. In this paper, we first demonstrates that the DNNs learning with label noise exhibits over-fitting issue on noisy labels because of the DNNs is too confidence in its learning capacity. More significantly, however, it also potentially suffers from under-learning on samples with clean labels. DNNs essentially should pay more attention on the clean samples rather than the noisy samples. Inspired by the sample-weighting strategy, we propose a meta-probability weighting (MPW) algorithm which re-weights the output probability of DNNs to prevent DNNs from over-fitting to label noise and alleviate the under-learning issue on the clean sample. MPW conducts an approximation optimization to adaptively learn the probability weights from data under the supervision of a small clean dataset, and achieves iterative optimization between probability weights and network parameters via meta-learning paradigm. The ablation studies substantiate the effectiveness of MPW to prevent the deep neural networks from overfitting to label noise and improve the learning capacity on clean samples. Furthermore, MPW achieves competitive performance with other state-of-the-art methods on both synthetic and real-world noises.

Deep learning-based methods have demonstrated significant breakthroughs in the application of hyperspectral image (HSI) classification. However, some challenging issues still exist, such as the overfitting problem caused by the limitation of training size with high-dimensional feature and the efficiency of spectral–spatial (SS) exploitation. Therefore, to efficiently model the relative position of samples within the generative adversarial network (GAN) setting, we proposed a dual-channel SS fusion capsule generative adversarial network (DcCapsGAN) for HSI classification. Dual channels (1-D-CapsGAN and 2-D-CapsGAN) are constructed by integrating the capsule network (CapsNet) with GAN for eliminating the mode collapse and gradient disappearance problem caused by traditional GAN. Meanwhile, octave convolution and multiscale convolution are integrated into the proposed model for further reducing the parameters of the CapsNet and extracting multiscale features. To further boost the classification performance, the SS channel fusion model is constructed to composite and switch the feature information of different channels, thereby facilitating the accuracy and robustness of the whole classification performance. Three commonly used HSI data sets are utilized to investigate the performance of the proposed DcCapsGAN model, and the performance of the experiment demonstrates that the proposed model can efficiently improve the classification accuracy and performance.

Arbitrary-oriented objects widely appear in natural scenes, aerial photographs, remote sensing images, etc., and thus arbitrary-oriented object detection has received considerable attention. Many current rotation detectors use plenty of anchors with different orientations to achieve spatial alignment with ground truth boxes. Intersection-over-Union (IoU) is then applied to sample the positive and negative candidates for training. However, we observe that the selected positive anchors cannot always ensure accurate detections after regression, while some negative samples can achieve accurate localization. It indicates that the quality assessment of anchors through IoU is not appropriate, and this further leads to inconsistency between classification confidence and localization accuracy. In this paper, we propose a dynamic anchor learning (DAL) method, which utilizes the newly defined matching degree to comprehensively evaluate the localization potential of the anchors and carries out a more efficient label assignment process. In this way, the detector can dynamically select high-quality anchors to achieve accurate object detection, and the divergence between classification and regression will be alleviated. With the newly introduced DAL, we can achieve superior detection performance for arbitrary-oriented objects with only a few horizontal preset anchors. Experimental results on three remote sensing datasets HRSC2016, DOTA, UCAS-AOD as well as a scene text dataset ICDAR 2015 show that our method achieves substantial improvement compared with the baseline model. Besides, our approach is also universal for object detection using horizontal bound box. The code and models are available at https://github.com/ming71/DAL.

Currently, reliable and accurate ship detection in optical remote sensing images is still challenging. Even the state-of-the-art convolutional neural network (CNN)-based methods cannot obtain very satisfactory results. To more accurately locate the ships in diverse orientations, some recent methods conduct the detection via the rotated bounding box. However, it further increases the difficulty of detection because an additional variable of ship orientation must be accurately predicted in the algorithm. In this article, a novel CNN-based ship-detection method is proposed by overcoming some common deficiencies of current CNN-based methods in ship detection. Specifically, to generate rotated region proposals, current methods have to predefine multioriented anchors and predict all unknown variables together in one regression process, limiting the quality of overall prediction. By contrast, we are able to predict the orientation and other variables independently, and yet more effectively, with a novel dual-branch regression network, based on the observation that the ship targets are nearly rotation-invariant in remote sensing images. Next, a shape-adaptive pooling method is proposed to overcome the limitation of a typical regular region of interest (ROI) pooling in extracting the features of the ships with various aspect ratios. Furthermore, we propose to incorporate multilevel features via the spatially variant adaptive pooling. This novel approach, called multilevel adaptive pooling, leads to a compact feature representation more qualified for the simultaneous ship classification and localization. Finally, a detailed ablation study performed on the proposed approaches is provided, along with some useful insights. Experimental results demonstrate the great superiority of the proposed method in ship detection.

Deep learning (DL) has become a hot topic in the research field of hyperspectral image (HSI) classification. However, with increasing depth and size of deep learning methods, its application in mobile and embedded vision applications has brought great challenges. In this article, we address a network architecture search (NAS)-guided lightweight spectral–spatial attention feature fusion network (LMAFN) for HSI classification. The overall architecture of the proposed network is guided by several conclusions of NAS, which achieves fewer parameters and lower computation cost with deeper network structure by exploiting multiscale Ghost grouped with efficient channel attention (ECA) module for adaptively adjusting the weights of different channels. It helps fully extract spectral–spatial discriminant features to avoid information loss of the dimension reduction operation. Specifically, a multilayer feature fusion method is proposed to extract the fusion information of the spectral–spatial features of each layer by considering complementary information of different hierarchical structures. Therefore, high-lever spectral–spatial attributes are gradually exploited along with the increase in layers and the fusion of layers. The experimental verification on three real HSI data sets demonstrates that the proposed framework presents more satisfying classification performance and efficiency with deeper network structure and lower parameter size.

Intelligent video surveillance is a vital technique in smart city construction, where detection of surveillance objects is generally achieved by subtracting estimated background from the raw video. Common wisdom of background estimation focuses on introducing meaningful structure or discriminative hypothesis to sparsity-based objectives. However, relaxation optimization, which is always considered a most effective solution, definitely leads to information loss. So, in this article, as to preserve more information, a new nonconvex sparsity model that can be solved directly by explicit solution is proposed for the stationary component of video. The solution, called generalized shrinkage thresholding operator, is designed by integrating the advantages of three common shrinkage operators. Then, for the regularly changing patterns, a purified dictionary learning operation is designed to find self-repeating texture patches. Eventually, foreground objects are detected by combining background subtraction with a spatiotemporal continuity constraint. Besides, built on optimizations of both models, we then show the way to refine the joint estimates using alternative optimization of all the subproblems. Experimental results have shown that, as to foreground detection task, when compared against current state-of-the-art techniques, the proposed model achieves comparable and often superior performance in terms of F-measure scores in most cases.

Deep learning models often use a flat softmax layer to classify samples after feature extraction in visual classification tasks. However, it is hard to make a single decision of finding the true label from massive classes. In this scenario, hierarchical classification is proved to be an effective solution and can be utilized to replace the softmax layer. A key issue of hierarchical classification is to construct a good label structure, which is very significant for classification performance. Several works have been proposed to address the issue, but they have some limitations and are almost designed heuristically. In this article, inspired by fuzzy rough set theory, we propose a deep fuzzy tree model which learns a better tree structure and classifiers for hierarchical classification with theory guarantee. Experimental results show the effectiveness and efficiency of the proposed model in various visual classification datasets.

Cartesian Genetic Programming (CGP) is a powerful and popular tool for automatic generation of computer programs to solve user defined tasks. This paper proposes a Co-evolutionary CGP (named Co-CGP) which can automatically gain high-order knowledge to accelerate the search. In the Co-CGP, two modules are working in cooperation to solve a given problem. One module focuses on solving a series of small scale problems of the same type to generate the building blocks. Simultaneously, the second module focuses on combing the available building blocks to construct the final solution. Besides, an adaptive control strategy is introduced to automatically evaluate the effectiveness of the building blocks and adjust the search behaviour adaptively so as to improve search efficiency. The proposed Co-CGP is tested on eight problems with different complexities. Experimental results show that the Co-CGP can significantly improve the performance of CGP, in terms of both search efficiency and accuracy.

In conventional wisdom of video modeling, background is often treated as the primary target and foreground is derived using the technique of background subtraction. Based on the observation that foreground and background are two sides of the same coin, we propose to treat them as peer unknown variables and formulate a joint estimation problem, called Hierarchical modeling and Alternating Optimization (HMAO). The motivation behind our hierarchical extensions of background and foreground models is to better incorporate a priori knowledge about the disparity between background and foreground. For background, we decompose it into temporally low-frequency and high-frequency components for the purpose of better characterizing the class of video with dynamic background; for foreground, we construct a Markov random field prior at a spatially low resolution as the pivot to facilitate noise-resilient refinement at higher resolutions. Built on hierarchical extensions of both models, we show how to successively refine their joint estimates under a unified framework known as alternating direction multipliers method. Experimental results have shown that our approach produces more discriminative background and demonstrates better robustness to noise than other competing methods. When compared against current state-of-the-art techniques, HMAO achieves at least comparable and often superior performance in terms of F-measure scores especially for video containing dynamic and complex background.

Product quantization (PQ) seems to have become the most efficient framework of performing approximate nearest neighbor (ANN) search for high-dimensional data. However, almost all existing PQ-based ANN techniques uniformly allocate precious bit budget to each subspace. This is not optimal, because data are often not evenly distributed among different subspaces. A better strategy is to achieve an improved balance between data distribution and bit budget within each subspace. Motivated by this observation, we propose to develop an optimized PQ (OPQ) technique, named distribution sensitive PQ (DSPQ) in this paper. The DSPQ dynamically analyzes and compares the data distribution based on a newly defined aggregate degree for high-dimensional data; whenever further optimization is feasible, resources such as memory and bits can be dynamically rearranged from one subspace to another. Our experimental results have shown that the strategy of bit rearrangement based on aggregate degree achieves modest improvements on most datasets. Moreover, our approach is orthogonal to the existing optimization strategy for PQ; therefore, it has been found that distribution sensitive OPQ can even outperform previous OPQ in the literature.

Background modeling is a critical component for various vision-based applications. Most traditional methods tend to be inefficient when solving large-scale problems. In this paper, we introduce sparse representation into the task of large-scale stable-background modeling, and reduce the video size by exploring its discriminative frames. A cyclic iteration process is then proposed to extract the background from the discriminative frame set. The two parts combine to form our sparse outlier iterative removal (SOIR) algorithm. The algorithm operates in tensor space to obey the natural data structure of videos. Experimental results show that a few discriminative frames determine the performance of the background extraction. Furthermore, SOIR can achieve high accuracy and high speed simultaneously when dealing with real video sequences. Thus, SOIR has an advantage in solving large-scale tasks.