Downsampling is essential in semantic segmentation for reducing computational cost and guiding the learning of class-discriminative features. Existing models typically rely on strided convolutions or patch splitting to obtain features with lower resolution. However, we observe that such operations often introduce edge jagging and texture degradation, the underlying cause is that aliasing of the high frequency induces phase distortion. We conducted a systematic analysis of phase distortion and identified two key properties: spatial non-uniformity (concentrated near boundaries) and directional sparsity (accumulated along a few dominant directions). These properties cause crucial high-frequency cues to be misrepresented or lost during sampling. To address this issue, we propose a frequency aware filter consisting of two complementary modules: a dynamic Gaussian kernel (DGK) and a learnable Gabor-based frequency selector (LFS). To mitigate spatial non-uniformity, the DGK predicts edge normals from gradients, applies strong low-pass filtering along the normal direction, and leaves the tangential direction virtually untouched, thereby suppressing phase distortion while preserving contour continuity. To handle directional sparsity, the Learnable Gabor Selector (LFS) then performs directional band-pass filtering to attenuate residual aliasing peaks and adaptively boost informative texture. We further introduce phase-error energy (PE) to quantify distortion severity. Visualization and quantitative results demonstrate that frequency-aware filter offers a plug-and-play remedy for aliasing, yielding sharper boundaries and consistent gains across datasets.

The objective of dehazing images is to enhance a hazy image by removing the atmospheric haze and restoring it to a clear, haze-free image. Previous studies often used simple deep learning methods combined with dehazing physical models, while neglecting the significance behind the parameters of the physical models. They did not employ appropriate deep learning methods to learn the parameters, which in turn hindered the effective restoration of the color and texture details in the images. To address the above issues, we propose a physical perception network (PPnet) for color and texture recovery. PPnet includes the following two major parts: the High-low frequency physical perception module(FPP) and the Attention-based physical estimation module(APE). The FPP module employ trainable filters to produce feature maps at both high and low frequencies, which are then applied to two identical single-parameter dehazing physical models. This module takes into greater consideration the contours and detailed features of images at different frequencies, and corrects the images through a physical model. Furthermore, in the physical model, we find that different parameters represent different physical meanings. Some parameters express global features while others represent local features. We propose the APE module, which utilized the pixel and channel attention method to simulate the physical parameter representation. Thus, it more accurately considers the physical significance of the parameters. PPnet exhibits a notable enhancement on the SOTS dataset the and HSTS dataset compared to existing state-of-the-art techniques.

In recent years, video stabilization methods have demonstrated significant achievements. However, learning-based video stabilization methods needs large paired unstable and stable video data, which is hard to collect. Besides, lots of online video stabilization methods always estimate trajectory by information of previous and several future frames, which is disadvantageous for devices that capture real-time video. In this paper, we propose Unsupervised Multi-stage Online Video Stabilization Network (UMOS-Net) to adapt trajectory with variable dynamic scenes by accurately perceiving the connections between current and previous frames, which does not reference future frames. Here, UMOS-Net is divided into two stages: trajectory estimation and trajectory smoothing. In trajectory estimation stage, considering the registration errors caused by blur or deformation in frames, such as fast-moving scenes, multi-scale content-aware modules are combined with adaptive optical flow loss to increase adaptability in the scenes with low overlap rates. Furthermore, lots of methods not only use several future frames, but use only the original camera trajectory as prior that accumulate from the first frame. The original camera trajectory is difficult to effectively smooth the current frame trajectory for cumulative error caused by low correlation of long-term frames. Therefore, a window-based dual-branch input recurrent neural network is designed in trajectory smoothing stage, in which stabilized historical trajectory is added as a supplement to get better trajectory smoothing results even in fast-moving scenes. Experiments show that our proposed UMOS-Net can obtain competitive results both qualitatively and quantitatively comparing to current representative online methods, especially in fast-moving scenes.

Recently, diffusion-based image compression has achieved significant progress in terms of rate–distortion-perception trade-off, these approaches have replaced decoders with conditional diffusion models to enhance the visual quality of reconstructed images. However, diffusion models introduce noise into the input image during the initial stages of the diffusion process, which may cause the potential degradation of crucial image information. To address these limitations, we propose a Hierarchical Conditional guidance Diffusion model for perceptual Image Compression (HCD-IC) to ensure the fidelity of reconstruction, in which hierarchical features with selected typical context provide informative guidance during the denoising process of diffusion model to preserve both structural integrity and fine details. Specifically, we design a Gated Scale-Cross module (GSC) to integrate and select representative semantics and details, which leverages a hierarchical feature interaction architecture and dynamic gated strategy to ensure more robust and expressive representations. Furthermore, we present a Conditional Control Diffusion decode module (CCD) to integrate time-step information and latent features augmented by GSC into the diffusion model, which can dynamically acquire the required time-aware conditional features during different denoise stages. Extensive experiments conducted on multiple public datasets demonstrate that our method outperforms state-of-the-art approaches in various quantitative realism metrics.

Conventional Self-Supervised Learning (SSL) exhibits notable limitations in fine-grained feature modeling due to pervasive issues in medical imaging, such as blurred organ boundaries, complex anatomical structures, and feature confusion caused by similar pathological patches, often leading to false positive sample interference. To address these challenges, this article proposes Gaze-Guided Semantic Contrastive Learning (G2Co), an innovative SSL algorithm inspired by visual diagnostic patterns of radiologists. At the semantic enhancement level, G2Co leverages a key information guidance mechanism to distinguish anatomical structures from background noise, thereby achieving fine-grained feature extraction. At the feature interaction level, G2Co introduces a cross-sample feature fusion strategy to extract discriminative features from potential positive samples, addressing feature confusion caused by visually similar patches. Furthermore, G2Co achieves refined modeling of tissue morphology and boundary characteristics by establishing inter-region mutual information maximization constraints. Finally, extensive experiments are conducted on the two widely used medical image datasets to demonstrate the effectiveness of our method.

Remote sensing change detection(RSCD) aims to identify the regions of interest that have changed between dual-temporal images. However, most deep models predict CD results by extracting multi-scale hybrid features, which could easily lead to ambiguous semantic boundaries; in addition, the existing feature acquisition tends to lack consideration of capturing their diversity usually causing poor model generalization. Thus, this paper decomposes the mixed features into change and invariant features jointly with stochastic distribution sampling and convolution thus accomplishing robust RSCD based on decoupled representations. In the training stage, the posterior distribution of the uncoupled features is first learned through label calibration to train the prior distribution generator; then, robust feature decoupling is implemented combining the convolutional feature separator with re-parameterized sampling over the decoupled posteriori distribution, and further aggregating the decoupled features through prototype learning; finally, the exceed-expectation loss regularizer is proposed to push or pull these positive and negative sample features to a more distant end, thereby increasing the inter-class distance by boosting the predicted expectation. In the testing stage, the robust RSCD based on decoupled representation is accomplished through the feature separator, decoupled prior distribution random sampling, and CD head without posterior distribution support. Experiments prove that DSFDcd has achieved remarkable results in terms of qualitative and quantitative metrics. Our codes will be available at https://github.com/iceking111/DSFDcd.

Zero-shot node classification is a vital task in the field of graph data processing, aiming to identify nodes of classes unseen during the training process. Prediction bias is one of the primary challenges in zero-shot node classification, referring to the model's propensity to misclassify nodes of unseen classes as seen classes. However, most methods introduce external knowledge to mitigate the bias, inadequately leveraging the inherent cluster information within the unlabeled nodes. To address this issue, we employ spectral analysis coupled with learnable class prototypes to discover the implicit cluster structures within the graph, providing a more comprehensive understanding of classes. In this paper, we propose a spectral approach for zero-shot node classification (SpeAr). Specifically, we establish an approximate relationship between minimizing the spectral contrastive loss and performing spectral decomposition on the graph, thereby enabling effective node characterization through loss minimization. Subsequently, the class prototypes are iteratively refined based on the learned node representations, initialized with the semantic vectors. Finally, extensive experiments verify the effectiveness of the SpeAr, which can further alleviate the bias problem.

Abstract Doctors typically use non‐contrast‐enhanced computed tomography (NCECT) in the treatment of kidney cancer to map kidney and tumour structural information to functional imaging single‐photon emission computed tomography, which is then used to assess patient kidney function and predict postoperative recovery. However, the assessment of kidney function and formulation of surgical plans is constrained by the low contrast of tumours in NCECT, which hinders the acquisition of accurate tumour boundaries. Therefore, this study designed a radiomics feature‐enhanced level‐set evolution (RFLSE) to precisely segment small‐sample low‐contrast kidney tumours. Integration of high‐dimensional radiomics features into the level‐set energy function enhances the edge detection capability of low‐contrast kidney tumours. The use of sensitive radiomics features to control the regional term parameters achieves adaptive adjustment of the curve evolution amplitude, improving the level‐set segmentation process. The experimental data used low‐contrast, limited‐sample tumours provided by hospitals, as well as the public datasets BUSI18 and KiTS19. Comparative results with advanced energy functionals and deep learning models demonstrate the precision and robustness of RFLSE segmentation. Additionally, the application value of RFLSE in assisting doctors with accurately marking tumours and generating high‐quality pseudo‐labels for deep learning datasets is demonstrated.

Remaining useful life (RUL) prediction links prognostic and predictive maintenance (PdM) decision-making. Since the prediction result serve as the basis for subsequent decision-making, it is vital to evaluate the performance of RUL prediction result. The most popular method tends to compares the actual RUL or run-to-failure data with the prediction result. However, in many real-world applications, such ground-truth measurements are unavailable during the prediction process. To address this issue, we propose a new method for real-time credibility evaluation of RUL prediction result in the absence of ground-truth measurements. The proposed method informs decision-makers of the confidence level of RUL prediction result prior to making a PdM decision. Multiple evaluation factors, such as accuracy, consistency, and effectiveness related to credibility, are proposed to utilize the information available during the prediction process based on the underlying stochastic process. Especially, a fuzzy comprehensive evaluation method is used to determine the credibility of RUL prediction result by comprehensively considering the evaluation factors, and an entropy weight method is used to update the weight of each factor. The rationality of the proposed credibility evaluation method is demonstrated using simulation and a benchmark dataset from NASA.

Data‐hunger is a persistent challenge in machine learning, particularly in the field of image processing based on convolutional neural networks (CNNs). This study systematically investigates the factors contributing to data‐hunger in machine‐learning‐based image‐processing algorithms. The results revealed that the proliferation of model parameters, the lack of interpretability, and the complexity of model structure are significant factors influencing data‐hunger. Based on these findings, this paper introduces a novel semi‐white‐box neural network model construction strategy. This approach effectively reduces the number of model parameters while enhancing the interpretability of model components. It accomplishes this by constraining uninterpretable processes within the model and leveraging prior knowledge of image processing for model. Rather than relying on a single all‐in‐one model, a semi‐white‐box model is composed of multiple smaller models, each responsible for extracting fundamental semantic features. The final output is derived from these features and prior knowledge. The proposed strategy holds the potential to substantially decrease data requirements under specific data source conditions while improving the interpretability of model components. Validation experiments are conducted on well‐established datasets, including MNIST, Fashion MNIST, CIFAR, and generated data. The results demonstrate the superiority of the semi‐white‐box strategy over the traditional all‐in‐one approach in terms of accuracy when trained with equivalent data volumes. Impressively, on the tested datasets, a simplified semi‐white‐box model achieves performance close to that of ResNet while utilizing a small number of parameters. Furthermore, the semi‐white‐box strategy offers improved interpretability and parameter reusability features that are challenging to achieve with the all‐in‐one approach. In conclusion, this paper contributes to mitigating data‐hunger challenges in machine‐learning‐based image processing through the introduction of a novel semi‐white‐box model construction strategy, backed by empirical evidence of its effectiveness.

Efficient mining machinery operation is essential for coal mining enterprises to reduce production costs, improve production efficiency, and maximize profits. As a key consumable part of mining machinery, the reliability of the pick directly determines the operational performance and service life of the machinery. During the mining process, the pick can be affected by natural wear and tear caused by the coal and rock and load shocks caused by the gangue and faults—that is, the result of the competing influence of soft failure caused by natural wear and tear and hard failure caused by random load shocks. At the same time, because the gangue and faults have a certain volume and hardness, a random load shock may last for a certain period of time, producing different acceleration effects on the pick wear under different hardness conditions. In this article, the reliability modeling of the pick competing failure process under random load shocks was studied by considering the influence of continuous shocks and the accelerated degradation at changing rates for different shock durations on the pick wear. First, the pick degradation model under random load shocks was established by considering the natural wear degradation, instantaneous shock degradation, and accelerated degradation at changing rates for different shock durations. Second, the pick reliability model under the competing failure mode was established on this basis. Finally, a numerical experiment and an effectiveness analysis of the pick reliability model were conducted based on the engineering data. The results showed that the competing failure reliability model of the pick considering continuous shocks and a changing degradation rate was more in line with engineering practice, which is of great significance for pick design optimization and improved reliability.

Compared to computed tomography (CT), magnetic resonance imaging (MRI) is more sensitive to acute ischemic stroke lesion. However, MRI is time-consuming, expensive, and susceptible to interference from metal implants. Generating MRI images from CT images can address the limitations of MRI. The key problem in the process is obtaining lesion information from CT. In this study, we propose a cross-modal image generation algorithm from CT to MRI for acute ischemic stroke by combining radiomics with generative adversarial networks. First, the lesion candidate region was obtained using radiomics, the radiomic features of the region were extracted, and the feature with the largest information gain was selected and visualized as a feature map. Then, the concatenation of the extracted feature map and the CT image was input in the generator. We added a residual module after the downsampling of the generator, following the general shape of U-Net, which can deepen the network without causing degradation problems. In addition, we introduced the lesion feature similarity loss function to focus the model on the similarity of the lesion. Through the subjective judgment of two experienced radiologists and using evaluation metrics, the results showed that the generated MRI images were very similar to the real MRI images. Moreover, the locations of the lesions were correct, and the shapes of lesions were similar to those of the real lesions, which can help doctors with timely diagnosis and treatment.

The fusion of multi-band images including far-infrared image (FIRI), near-infrared image (NIRI), and visible image (VISI) primarily suffers from four challenges. One is the problem of simultaneous fusion for multiple images. Most existing methods are oriented towards the fusion of two objects, which is generally achieved with a sequential fusion method. This means that intermediate fusion results are repeatedly integrated with the unprocessed images until all images have been fused. However, this may amplify the blurring effect, and even engender artifacts. Second, consistent training labels for image fusion cannot currently be obtained for some types of images (e.g., medical images, and multi-band images), which may lead to the failed application of supervised learning methods. Third, the existing methods often do not directly focus on the potential mapping relationship between the original, and resulting images, which usually increases the unpredictability of the fusion results. Fourth, redundant features or singularities are often not eliminated in the general fusion process, and both may interfere with or even obscure significant features in the source images. To address the abovementioned problems, this paper proposes a latent representation learning model that can synchronously integrate multi-band images without samples. Specifically, the model can capture the clean, and distinctive features of the originals via latent low-rank, and sparse embedding. The extracted intrinsic features are projected onto the target fusion space through an assumed mapping relationship. The final results were obtained through the designed optimization algorithm. In addition, numerous experiments were implemented to prove the rationality, and feasibility of the proposed fusion model with subjective evaluation, objective indexes, and convergence analysis.

Aiming at ash and slag of boiler heated surface will reduce the heat transfer efficiency and security. This paper adopts clearness factor as the indicator to monitor healthy condition of the boiler heated surface, and put forward a model combining empirical mode decomposition (EMD) and long short-term memory (LSTM) model to predict boiler ash accumulation in the future. EMD can decompose the time series into a series of frequency-stable intrinsic mode functions. In addition, the special gate structure inside LSTM makes it possible to mine the long-term dependencies in the time series. The combination of the two increases the prediction accuracy of the time series. It is verified by simulation software that the model has a satisfactory accuracy in predicting healthy condition of the boiler heated surface, and the feasibility and effectiveness of the model are verified.

Ultrasonography is one of the main imaging methods for diagnosing thyroid nodules. Automatic differentiation between benign and malignant nodules in ultrasound images can greatly assist inexperienced clinicians in their diagnosis. The key of problem is the effective utilization of the features of ultrasound images. In this study, we propose a method that is based on the combination of conventional ultrasound and ultrasound elasticity images based on a convolutional neural network and introduces richer feature information for the classification of benign and malignant thyroid nodules. First, the conventional network model performs pretraining on ImageNet and transfers the feature parameters to the ultrasound image domain by transfer learning so that depth features may be extracted and small samples may be processed. Then, we combine the depth features of conventional ultrasound and ultrasound elasticity images to form a hybrid feature space. Finally, the classification is completed on the hybrid feature space, and an end-to-end CNN model is implemented. The experimental results demonstrate that the accuracy of the proposed method is 0.9470, which is better than that of other single data-source methods under the same conditions.

Model management plays an essential role in surrogate-assisted evolutionary optimization of expensive problems, since the strategy for selecting individuals for fitness evaluation using the real objective function has substantial influences on the final performance. Among many others, infill criterion driven Gaussian process (GP)-assisted evolutionary algorithms have been demonstrated competitive for optimization of problems with up to 50 decision variables. In this paper, a multiobjective infill criterion (MIC) that considers the approximated fitness and the approximation uncertainty as two objectives is proposed for a GP-assisted social learning particle swarm optimization algorithm. The MIC uses nondominated sorting for model management, thereby avoiding combining the approximated fitness and the approximation uncertainty into a scalar function, which is shown to be particularly important for high-dimensional problems, where the estimated uncertainty becomes less reliable. Empirical studies on 50-D and 100-D benchmark problems and a synthetic problem constructed from four real-world optimization problems demonstrate that the proposed MIC is more effective than existing scalar infill criteria for GP-assisted optimization given a limited computational budget.

When using a fixed number of neighbors for training a local surrogate model in surrogate assisted evolutionary optimization algorithms, it may suffer from the large uncertainty because the actual distribution of candidate's neighborhood may be neglected. In this paper, we propose to firstly analyze the distribution characteristics of candidate's neighborhood through a modified overlapping clustering method before training a local surrogate, and then use the clustering based evolution control strategy or model management strategy to facilitate the evolutionary algorithm to converge to the right optimum. Simulation results on four widely used benchmark functions demonstrate the efficacy of the proposed method.

Surrogate models have shown to be effective in assisting metaheuristic algorithms for solving computationally expensive complex optimization problems. The effectiveness of existing surrogate-assisted metaheuristic algorithms, however, has only been verified on low-dimensional optimization problems. In this paper, a surrogate-assisted cooperative swarm optimization algorithm is proposed, in which a surrogate-assisted particle swarm optimization (PSO) algorithm and a surrogate-assisted social learning-based PSO (SL-PSO) algorithm cooperatively search for the global optimum. The cooperation between the PSO and the SL-PSO consists of two aspects. First, they share promising solutions evaluated by the real fitness function. Second, the SL-PSO focuses on exploration while the PSO concentrates on local search. Empirical studies on six 50-D and six 100-D benchmark problems demonstrate that the proposed algorithm is able to find high-quality solutions for high-dimensional problems on a limited computational budget.

As a population-based meta-heuristic technique for global search, particle swarm optimization (PSO) performs quite well on a variety of problems. However, the requirement on a large number of fitness evaluations poses an obstacle for the PSO algorithm to be applied to solve complex optimization problems with computationally expensive objective functions. This paper extends a fitness estimation strategy for PSO (FESPSO) based on its search dynamics to reduce fitness evaluations using the real fitness function. In order to further save the fitness evaluations and improve the estimation accuracy, a similarity measure and a reliability measure are introduced into the FESPSO. The similarity measure is used to judge whether the fitness of a particle will be estimated or evaluated using the real fitness function, and the reliability measure is adopted to determine whether the approximated value will be trusted. Experimental results on six commonly used benchmark problems show the effectiveness and competitiveness of our proposed algorithm. Preliminary empirical analysis of the search behavior is also performed to illustrate the benefit of the proposed estimation mechanism.

Collaborative model and process model for complex machinery products collaborative design are studied in this paper. The authors propose a layer collaborative model, which is used to describe the layer collaborative process in complex machinery design firstly. It contains design layer dimension, design objective dimension, design cycle/period dimension and design constraint dimension. The model is described by state-space method and built up by object layer dividing strategy. Then the authors present a process model for collaborative design and definite it in formalization, which is used to describe design process of each layer in collaborative model. At the same time activity type, activity state, activity operation and activity process in collaborative design are analyzed. It affords information basis to organize complex machinery products collaborative design orderly on web by building the multi-layer collaborative model.

Estimation of Distribution Algorithm (EDA) is a novel evolutionary computation, which mainly depends on learning and sampling mechanisms to manipulate the evolutionary search, and has been proved a potential technique for complex problems. However, EDA generally spend too much time on the learning about the probability distribution of the promising individuals. The paper propose an improved EDA based on copula theory (copula-EDA) to enhance the learning efficiency, which models and samples the joint probability function by selecting a proper copula and learning the marginal probability distributions of the promising population. The simulating results prove the approach is easy to implement and is validated on several problems.

Due to the design knowledge discrepancy during collaborative design, conflicts can be revealed from the process of collaborative design decision. A critical element of collaborative design would be conflict resolution. The conflict resolution is correlative with both of the knowledge granulation and specific method provided. In this paper, granularity is used to describe rules acquainted based on the concept of rough degree. In this condition, A feasible distance formula between different rules is constructed according to the requirement of describing the distance between different rules and the definition of attribute significance in information system with rough theory, in which the operator of attribute importance is introduced, and then traditional conflict resolutions are analyzed. As a result, the concept of rule set evolution is put forward, and the instance is introduced to explain the efficiency of new method.

The paper develops an adaptive particle swarm optimization(PSO) based on its simplification. With the hope to prompt its global optimization performance, the improved menthod introduces a PD controller into the architecture of the standard PSO(SPSO-PD). The PD controller can control the particle dynamics, and prompt the particles to respond to the change of their exemplars correctly and rapidly, further to surpass the limit of their exemplars with more chances, which greatly contributes to a successful global search. The proposed SPSO-PD was applied to some well-known benchmarks and compared with the standard PSO. The relative experimental results show SPSO-PD performs better than SPSO on the complex optimization functions.

In standard particle swarm optimization (PSO), the velocity only provides a position displacement contrast with the longer computational time. To avoid premature convergence, a new modified PSO is proposed in which the velocity considered as a predictor, while the position considered as a corrector. The algorithm gives some balance between global and local search capability, and results the high computational efficiency. The optimization computing of some examples is made to show the new algorithm has better global search capacity and rapid convergence rate.

Particle swarm optimizer is a novel algorithm where a population of candidate problem solution vectors evolves "social" norms by being influenced by their topological neighbors. The standard particle swarm optimizer (PSO) may prematurely converge on suboptimal solutions that are not even guaranteed to be local extrema. A new particle swarm optimizer, called stochastic PSO (SPSO), which combined with tabu technique is presented based on the analysis of the standard PSO. And because of its local search capability, the SPSO is more efficient. And the global convergence analysis is made using the F. Solis and R. Wets' research results. Finally, several examples are simulated to show that SPSO is more efficient than the standard PSO.

Particle swarm optimizer (PSO) is a new evolutionary computation method, which has been successfully applied to many fields. Through mechanism analysis of the standard particle swarm optimizer, a linear equivalent representation of the velocity update equation is given. Thus a new particle swarm optimizer-nonlinear particle swarm optimizer (NPSO) is described in this paper. It can dynamically adjust the limit position that distribute within the ellipse located by the best positions of the population and itself, and drop out of the local minimum point. The simulation results show the correctness and efficiency of the presented methods.

Through mechanism analysis of genetic algorithm (GAs), every genetic operator of SGA and their combination action can be considered as linear transformations to the corresponding individuals. By modifying the traditional genetic operators, the nonlinear genetic algorithm (NGA) is introduced. In this paper, abstract model of NGA based on function group is discussed in which every function is selected with some probability, and if the function group is correctly selected, then the algorithm can be convergent to global optimal within every given generation number. With this technique the premature convergence and calculation efficiency may be solved.

Through mechanism analysis of simple genetic algorithm (SGA) every genetic operator can be considered as a linear transform. So some disadvantages of SGA may be solved if genetic operators are modified to nonlinear transforms. According to the above method, nonlinear genetic algorithm is introduced, and different nonlinear genetic operators with some probability are designed and applied to numerical optimization problems. The optimization computing of some examples is made to show that the new genetic algorithm is useful and simple.

In this paper, we proposed a multi-agent system for planning and scheduling of the integrated steel processes, which is an extremely complex task requiring the consideration of numerous constraints and objectives. The system is implemented using the agent based Asynchronous Team (A-Team) architecture in which multiple solution methods cooperate by evolving shared population of solution. Each process in the integrated production environment is assigned to an agent, which performs, independently its scheduling at a local level by using its sub-agent: construction agents, improvement agents and destruction agents.