Iterative Strategy Research Articles

Joint spatio-temporal modeling for visual tracking

Similarity-based approaches have made significant progress in visual object tracking (VOT). Although these methods work well in simple scenes, they ignore the continuous spatio-temporal connection of the object in the video sequence. For this reason, tracking by spatial matching solely can lead to tracking failures because of distractors and occlusion. In this paper, we propose a spatio-temporal joint-modeling tracker named STTrack which implicitly builds continuous connections between the temporal and spatial aspects of the sequence. Specifically, we first design a time-sequence iteration strategy (TSIS) to concentrate on the temporal connection of the object in the video sequence. Then, we propose a novel spatial temporal interaction Transformer network (STIN) to capture the spatio-temporal correlation of the object between frames. The proposed STIN module is robust in object occlusion because it explores the dynamic state change dependencies of the object. Finally, we introduce a spatio-temporal query to suppress distractors by iteratively propagating the target prior. Extensive experiments on six tracking benchmark datasets demonstrate that the proposed STTrack achieves excellent performance while operating in real-time. The code is publicly available at https://github.com/nubsym/STTrack.

Usability and feasibility of E-nergEYEze: a blended vision-specific E-health based cognitive behavioral therapy and self-management intervention to reduce fatigue in adults with visual impairment

BackgroundOver 50% of adults with visual impairment experience severe fatigue. Therefore, we developed a guided E-health intervention based on cognitive behavioral therapy and self-management to reduce fatigue in this population. This pilot study evaluated the usability, feasibility, fidelity and potential effectiveness of E-nergEYEze.MethodsE-nergEYEze was developed by a design team and customized by conducting a pilot study using an iterative development strategy. The intervention was first tested in a usability study among adults with visual impairment (n = 5). Participants were asked to think-aloud while exploring the intervention features and a semi-structured interview was performed afterwards. Subsequently, the enhanced intervention was tested in a feasibility study. Adults with visual impairment and severe fatigue (n = 10) followed the intervention partially with guidance from a social worker and one-time computer trainer support. Fatigue severity (Checklist Individual Strength), fatigue impact (Modified Fatigue Impact Scale) and cognitive behavioral therapy skills (Competencies of Cognitive Therapy Scale-Self Report) were measured at baseline and at three months follow-up and analyzed with the Wilcoxon signed-rank test. The intervention was evaluated through evaluation forms.ResultsThe usability study resulted in adjustments to content and lay-out with regard to optically shortened text sentences, separate pages for information and assignments with one read-aloud audio and an additional descriptive explanation of page content. Digital challenges were overcome with mandatory computer training and e-platform modifications. The feasibility study showed a positive trend in reducing fatigue severity (Z -6.108; P < .001; SD 8.4), impact of fatigue (Z − 4.451; P < .001; SD 11.4) and cognitive behavioral therapy skills (Z -2.278; P = .023; SD 19.3). Participants gave useful feedback regarding accessibility, content and guidance, with an overall positive experience. The intervention was rated with a median score of 8 (range 7–10).ConclusionWe developed, evaluated and optimized E-nergEYEze by applying a user-centered and iterative approach. E-nergEYEze showed a promising trend to reduce fatigue severity and impact of fatigue and to increase cognitive behavioral therapy skills. The study methods were feasible and the fidelity of the intervention protocol was suitable. Performing a randomized controlled trial is warranted to give insight into whether E-nergEYEze is cost-effective in reducing severe fatigue in adults with visual impairment.Trial registrationInternational Clinical Trial Registry Platform: NL7764. Date registered: 28-05-2019.

Combination Therapy Use and Associated Events in Clinical Practice Following Dissemination of Trial Findings: A De-Implementation Study Using Interrupted Time Series Analysis

PurposeIt takes 17 years, on average, for trial results to be implemented into practice. Using data from the Department of Veterans Affairs (VA), this study assessed the potential impact on clinical practice of the dissemination of findings from a randomized, controlled trial reporting harm with the use of combination therapy. Communication between research and VA Pharmacy Benefits Management Services (PBM) provided the impetus for communication from the PBM about the findings of the trial in accordance with policy. MethodsIn this de-implementation study, interrupted time series analysis was used for assessing prescribing patterns and adverse clinical events before and after the dissemination of the trial findings. The de-implementation strategy was multicomponent and multilevel. Strategies were aligned with categories outlined in the Expert Recommendations for Implementing Change: train and educate stakeholders, use evaluative and iterative strategies, develop stakeholder inter-relationships, change infrastructure, provide interactive assistance, and engage consumers. VA patients with type 2 diabetes mellitus, chronic kidney disease stages 1 to 3, and a moderate or severe albuminuria who received care between July 2008 and November 2017 were included. Patients were subgrouped according to treatment with an angiotensin-converting enzyme inhibitor + angiotensin receptor blocker. The primary end point was the prevalence of combination therapy use. Secondary end points were the incidences of acute kidney injury and hyperkalemia. FindingsThis study followed 712,245 patients, 9297 of whom used combination therapy. Data were available from 428,535 and 283,710 patients pre- and post-intervention, respectively; among these, 8324 and 973 patients used combination therapy, the median ages were 66 and 68 years, and 96.92% and 98.82% were men. One month following communication from the PBM, the reductions in combination therapy users, acute kidney injury events, and hyperkalemia were 331.94 (95% CI, 500.27–163.32), 36.58% (95% CI, 31.90%–41.95%), and 25.49% (95% CI, 14.17%–36.07%) per 100,000 patients per month, respectively (all, P < 0.001), whereas before the communication, these changes were +14.84 (95% CI, 10.27–19.42), –3.46% (95% CI, 3.18–3.74), and –3.27% (95% CI, 2.66%–3.87%) (all, P < 0.001). ImplicationsThe apparent speed and impact of the implementation of changes resulting from the dissemination of trial findings into VA clinical practice are encouraging. The speed of implementation was much faster than average for health care providers in the United States. Established communications between research and clinical practice, as well as established policy and communications between PBM and clinical practice, may be a model for other health care organizations.

An Elastic-Window-Based Method for the Underdetermined Problem in Linear Spectral Unmixing to Enhance the Spatial Resolution of the Normalized Difference Vegetation Index Time Series

Inverting land cover reflectance or derived indices from low-spatial-resolution images to refine the spatial resolution of this data is cost-effective for land surface monitoring applications that face technical or budget limitations. Based on the linear spectral mixing model, many approaches have successfully unmixed coarse mixed pixels using high-spatial-resolution land cover maps in the past decades. However, in some cases, the solutions of linear systems composed of several mixed pixels may not be acquired due to the underdetermined problem. This study presents the causes of this problem and proposes an iterative approximation strategy to address it. An elastic-window-based algorithm was developed, where the initial size of the window was calculated based on the land cover of the mixed pixel. Mixed pixels of neighborhoods with similar land covers were then selected to form the unmixing linear system, which was examined through a simulation test to ensure it was not underdetermined. Otherwise, the window would expand to include more adjacent pixels. This process was repeated until a successful solution was obtained. A statistical analysis of sixty land cover maps from around the globe shows that the underdetermined problem exists at a low level but becomes more serious with an increase in mixed scale. The results demonstrate that the proposed algorithm effectively prevents the underdetermined problem for mixed pixels of different scales and can be integrated into the coarse NDVI downscaling procedure to refine spatial resolution. This study provides a reference for estimating underdetermined mixed pixels and benefits applications that require dealing with the inversion of land cover values directly.

Fault detection and identification for rolling mill main drive system based on integrated observer under iterative learning strategy

In this article, the problem of multiple fault detection, isolation and reconfiguration of the rolling mill main drive system containing external disturbances is investigated. Considering the nonlinear frictional damping between the rolls and the rolled parts, a nonlinear mathematical model of the main drive system of the mill is established. A comprehensive fault diagnosis scheme based on observer is addressed for this system subjected to unknown external interference. The proposed scheme is divided into two parts. In the first stage, a set of sliding mode observers is designed for system fault detection, and a fault isolation criterion is proposed based on observer redundancy and generalised residual set theory to reveal the fault source. In the second stage, combined with the iterative learning algorithm, an iterative learning-unknown input observer is constructed to realise the accurate estimation of the fault signal. Unlike the existing fault estimation methods, the iterative learning-unknown input observer designed in this article uses the state estimation error of the previous iteration to estimate the fault signal in the current iteration period. Using [Formula: see text] synthesis to design observers for the system will guarantee fault diagnosis robustness. The Lyapunov theory and linear matrix inequality are introduced to prove the convergence of the proposed observer. The simulation study of a 1780-mm hot strip mill evaluates the proposed scheme. Simulation results demonstrate that the sliding mode observer approach can detect faults in the main drive system and isolate faults accurately. In contrast, the iterative learning-unknown input observer method has the lowest fault reconfiguration error (99.87% smaller than the extended state observer, 99.77% smaller than the unknown input observer) and achieves accurate fault signal tracking.

Implementing a Structured Health Care Transition (HCT) Process for Sickle Cell Disease Using Quality Improvement: Results from the Sickle Cell Trevor Thompson Transition Project HCT Learning Collaborative

Ensuring consistent healthcare during transition is challenging for sickle cell disease (SCD). Emerging adults with SCD (EASCD) experience triple the mortality rates and double the acute care utilization rates compared to younger and older ages. The transition period is also plagued by reduced quality of life (QOL) with increased missed school days per year and worse standardized test scores. The Sickle Cell Trevor Thompson Transition (ST3P-UP) study aimed to improve acute care utilization and QOL for EASCD by standardizing the transition process while engaging the community as an equal partner. A prior regional stakeholder needs assessment identified care standardization, establishing individual emergency care plans (ECPs), and community engagement as critical success factors for improved access to and engagement in care during transition. The ST3P-UP SCD health hare transition (HCT) learning collaborative engaged 14 pediatric and adult paired sites that included a partner community-based organization (CBO) using the concept of a site “Triad” as the critical framework of each sites transition team. The collaborative aim statement established at a kickoff meeting was “ to improve the health of EASCD by implementing an education-based transition program in pediatric and adult SCD clinics to be accomplished by 24 months, using the Model for Improvement and the Got Transition Six Core Elements of Transition (6CE) model. ” Using the 6CE framework established by Got Transition® and the Model for Improvement QI methodology, we established a SCD HCT Process Measurement Tool (PMT) and 2 outcome measures for the transition learning collaborative namely: transition readiness assessment (RA), and an individualized emergency care plan (ECP). Monthly virtual meetings provided coaching on HCT recommendations, use of Plan-Do-Study-Act cycles, and practical guidance on implementing the 6CE. Sites were given a change package that included provider and patient education materials, resources to establish and strengthen a QI team, transition billing and coding resources etc.) to improve their transition processes. Implementation of 6CE was measured across 14 clinical sites, each comprised of both pediatric and adult clinics, who cared for 1625 EASCD aged 16-25. Clinical programs varied: 12 urban, 2 rural; 12 academic, 2 non-academic; 6 small, 8 large. The HCT-PMT assessed 6CE implementation adherence using iterative QI strategies at baseline (2018) and every six months thereafter for 54mo (2022). Pre-post results were compared for the overall group and by practice size. All 14 sites made substantial progress towards implementing a structured HCT process within 24 months. Overall HCT-PMT scores improved by 402% by the end of the 54mo collaborative (19.4 ± 14.1 to 97.4 ± 5.2). PMT scores improved significantly by 312% (24 to 99) for pediatrics and 539% (15 to 96 for adult sites. Pediatric PMT scores were higher at each assessment interval vs. adult scores, however adult sites achieved a higher degree of improvement. Smaller sites achieved a higher degree of improvement vs. large sites. Large site PMT scores improved 362% (21 to 97); small site PMT scores improved 444% (18 to 98). Additionally, sites demonstrated substantial improvements with having current ECPs for EASCD annually (completed and reviewed within the past year). At baseline, only 20% of pediatric and 3% of adult EASCD had a current ECP; only 38% of pediatric and 20% of adult EASCD had a current RA. Run charts demonstrated significant improvement with implementing the ECP measure. Pediatric sites had significant and sustained improvement in adherence to the ECP measure for &amp;gt;6mo. Adult sites started with a lower baseline score. Although they met their goal of showing process capability, they have not yet demonstrated sustained improvement. By Dec. 2022, site adherence with current ECP improved to 84% and 70% for pediatric and adult sites respectively. Adherence with current RA also improved to 76% and 70% for pediatric and adult sites respectively. All 14 sites showed significant improvement with implementing a structured HCT process aligned with the 6CE using QI confirming its feasibility in establishing a standardized and measurable transition process that can be applied to various chronic conditions.

Railway catenary optimisation with span variability by an iterative optimisation algorithm for large numbers of parameters

Optimising the catenary topology to improve the pantograph–catenary interaction is computationally expensive. Especially in the case of large numbers of parameters. Unfortunately, varying span lengths and dropper distribution in realistic systems requires the inclusion of a vast number of parameters, making the optimisation problem extremely complicated. Published optimisation works assume that the catenary is a perfectly periodical structure to reduce the number of parameters involved in the optimisation and keep the computational cost below a feasible limit. This paper attempts to deal with the scientific challenge of optimising a realistic system by proposing an iterative optimisation strategy consisting of optimising groups of parameters separately and repeating the optimisations iteratively to consider the dependence between parameters that otherwise would need to be optimised together. This solution performs topological catenary optimisation in catenary zones with span variability or transition zones in which optimisations that assume periodicity are no longer valid.

Fast flow prediction of airfoil dynamic stall based on Fourier neural operator

Dynamic stall on airfoil is of great importance in engineering applications. In the present work, Fourier neural operator (FNO) is applied to predict flow fields during the dynamic stall process of the NACA0012 airfoil. Two cases with different angles of attack are simulated by Reynolds averaged numerical simulation with the Spalart–Allmaras (SA) model at Re=4×104. A prediction model is directly constructed between the flow fields at several previous time nodes and that at the future time node by FNO. The prediction of sequence flow fields based on the iterative prediction strategy is achieved for the dynamic stall. The results show that FNO can achieve a fast and accurate prediction of streamwise velocity, normal velocity, pressure, and vorticity for both cases. The dynamics of vortices around the airfoil is analyzed to demonstrate the prediction accuracy of FNO. In addition, FNOs with different configurations are tested to achieve a lower error and a shorter training time-consuming.

SEBP_HNHC: Stacking Ensemble-based Bi-level Predictor for Human Non-Histone Crotonylation combining with iterative feature representation strategy

Accurate identification of crotonylation on human non-histone sites plays a crucial role in human biological research. Compared with traditional experimental methods that are labor-intensive and time-consuming, computational methods are more and more popular in recent years. Most of previous research methods are only for the prediction of crotonylation sites of histones, and there is a lack of targeted prediction tools for the identification of crotonylation sites of human non-histone proteins. In this study, we propose a stacking ensemble-based bi-level predictor for human non-histone crotonylation combining with iterative feature representation strategy, named SEBP_HNHC. To take full advantage of all the training information and solve the extreme imbalance between positive samples and negative samples, we propose a two-layer training data structure including preliminary training dataset and balanced training dataset. Preliminary training dataset is firstly divided into twelve subset using a special data division method. Then, multi-view feature encoding schemes are integrated to comprehensively represent the attribution of samples in each training subset. Next, Gini index combined with sequential forward search is employed for feature optimization for each preliminary training subset. Then, 84 preliminary models are built based on seven kinds of classifiers for twelve preliminary training subset that are now more commonly used and efficient. And then, a balanced training dataset is utilized to synthesize the capabilities of 84 preliminary models by generating 84 probability values as 84-D features for stacking ensemble. Finally, deep ensemble features are mined through iterative feature representation strategy. The experiment results show that the SEBP_HNHC method achieves robust performance in crotonylation sites prediction. Therefore, SEBP_HNHC is an effective tool for identification of crotonylation on lysine sites in human non-histone proteins.

A Quicker Iteration Method for Approximating the Fixed Point of Generalized α-Reich-Suzuki Nonexpansive Mappings with Applications

Fixed point theory is a branch of mathematics that studies solutions that remain unchanged under a given transformation or operator, and it has numerous applications in fields such as mathematics, economics, computer science, engineering, and physics. In the present article, we offer a quicker iteration technique, the D** iteration technique, for approximating fixed points in generalized α-nonexpansive mappings and nearly contracted mappings. In uniformly convex Banach spaces, we develop weak and strong convergence results for the D** iteration approach to the fixed points of generalized α-nonexpansive mappings. In order to demonstrate the effectiveness of our recommended iteration strategy, we provide comprehensive analytical, numerical, and graphical explanations. Here, we also demonstrate the stability consequences of the new iteration technique. We approximately solve a fractional Volterra–Fredholm integro-differential problem as an application of our major findings. Our findings amend and expand upon some previously published results.

Enhanced robust multimode process monitoring under dirty data via difference-based decomposition of matrix

Traditional data-driven methods generally suppose the training dataset is not corrupted by outliers. However, outliers are inevitable in the real industrial processes even with a relatively high ratio, which degrades the accuracy of data-based models. For multimode process monitoring, outliers may deteriorate the accuracy of both mode identification and fault detection. However, the existing robust methods can hardly deal with a large percentage of outliers, i.e., dirty data in neither single nor multimode processes. In this paper, a robust multimode process monitoring scheme is developed by alleviating the negative effect of dirty data. A difference-based decomposition of matrix (DDM) algorithm is first proposed to divide the data into a basic subpart and a non-basic subpart. The optimization function of the proposed DDM algorithm is solved by the alternating direction method of multipliers (ADMM). In the off-line procedure, an iterative decomposition strategy is designed based on the proposed DDM approach to identify dirty data and obtain the clean dataset of each mode. In the on-line procedure, an on-line sample identification strategy is developed by the type indicator derived from the DDM algorithm to determine whether the current sample belongs to a mode, the dirty data, or the fault. A numerical example and an industrial-scale multiphase flow facility indicate the proposed method can improve the accuracy of both mode identification and fault detection for multimode processes even under dirty data.

Synchronous disturbance suppression of active magnetic bearing rotor systems using variable period adaptive control algorithm

This paper aims to investigate the issue of suppressing synchronous vibration in active magnetic bearing (AMB) systems. The rotor mass imbalance is a challenging problem for the AMB system, which will cause synchronous vibration and severely affect the system stability and safety. However, existing algorithms based on the iterative feed-forward control strategy find it difficult to simultaneously ensure system’s stability and competent convergence performance. They struggle to maintain tolerable robustness within the system operating bandwidth. This paper proposes a variable period adaptive control algorithm to address these challenges. The overall control scheme can achieve current and displacement elimination separately by switching control structures and is applicable to variable speed conditions. This algorithm can adaptively adjust the iteration period based on rotational speed to ensure both splendid convergence performance and system dynamic performance. Furthermore, this work employs the Lyapunov method to analyze the algorithm’s asymptotic stability and provides a lookup table for algorithm parameters. This could significantly enhance the robustness and stability of the algorithm within the system operating bandwidth. Finally, this methodology has proven to be effective and superior in simulations and experiments.

An iterative output feedback controller-design strategy for energy-to-peak structural vibration control

ABSTRACT To address the issue of inadequate control effect resulting from the challenge of measuring state feedback information comprehensively, this paper proposes an iterative output feedback energy-to-peak control method that solely depends on partial external observation data. This paper presents the computational steps for achieving a suboptimal solution to the output feedback bilinear matrix inequality constrained problem using an iterative linear matrix inequality approach. The proposed method overcomes the challenge of numerically solving the optimization problem. The unknown variables in the output feedback linear matrix inequality optimization problem are transformed equivalently to obtain a new linear matrix inequality optimization problem. The numerical results of the two optimization problems are then used to iterate with each other and obtain a suboptimal solution. A comprehensive analysis was conducted on an 8-story building structure, utilizing state feedback, output feedback, and iterative output feedback energy-to-peak control techniques. The results of the study indicate that the proposed iterative output feedback energy-to-peak control approach efficiently reduces the displacement and acceleration response of the structure, exhibiting superior control effectiveness than both the state feedback and output feedback energy-to-peak control methods. The findings demonstrate the practicality and versatility of the suggested method.

Black-box Attack against Self-supervised Video Object Segmentation Models with Contrastive Loss

Deep learning models have been proven to be susceptible to malicious adversarial attacks, which manipulate input images to deceive the model into making erroneous decisions. Consequently, the threat posed to these models serves as a poignant reminder of the necessity to focus on the model security of object segmentation algorithms based on deep learning. However, the current landscape of research on adversarial attacks primarily centers around static images, resulting in a dearth of studies on adversarial attacks targeting Video Object Segmentation (VOS) models. Given that a majority of self-supervised VOS models rely on affinity matrices to learn feature representations of video sequences and achieve robust pixel correspondence, our investigation has delved into the impact of adversarial attacks on self-supervised VOS models. In response, we propose an innovative black-box attack method incorporating contrastive loss. This method induces segmentation errors in the model through perturbations in the feature space and the application of a pixel-level loss function. Diverging from conventional gradient-based attack techniques, we adopt an iterative black-box attack strategy that incorporates contrastive loss across the current frame, any two consecutive frames, and multiple frames. Through extensive experimentation conducted on the DAVIS 2016 and DAVIS 2017 datasets using three self-supervised VOS models and one unsupervised VOS model, we unequivocally demonstrate the potent attack efficiency of the black-box approach. Remarkably, the J&amp;F metric value experiences a significant decline of up to 50.08% post-attack.

An iterative augmented unscented Kalman filter for simultaneous state-parameter-input estimation for systems with/without direct feedthrough

A few unscented Kalman filters (UKFs) have been developed for simultaneous state-parameter-input estimation, however, these UKFs often have at least one of these limitations: requiring ad hoc procedures; requiring displacement measurements, which may be inconvenient to obtain, to fuse with acceleration measurements to avoid the “drift phenomenon” in the estimated displacement responses and unknown inputs; requiring a (full-rank) feedthrough matrix; and undesirable computational efficiency. To overcome these limitations, a novel iterative augmented unscented Kalman filter (IAUKF) for simultaneous state-parameter-input estimation for systems either with or without direct feedthrough is developed and examined in this study. The IAUKF closely follows the Bayes’ theorem with no ad hoc procedure, featuring theoretical simplicity. An iterative strategy is introduced to improve the accuracy of the unknown input estimation at each time step, so that the “drift phenomenon” encountered by the existing augmented Kalman filter is avoided. Further, the IAUKF requires no prior models for the time histories of the unknown inputs and applies to systems with/without direct feedthrough. With realistic assumptions, the steps for computationally efficient implementations of the IAUKF are proposed. Numerical examples of nonlinear structural systems with/without direct feedthrough of the unknown inputs are adopted to examine the effectiveness and the computational efficiency of the IAUKF.

A resource of single-cell gene expression profiles in a planarian Dugesia japonica.

The freshwater planarian Dugesia japonica maintains an abundant heterogeneous cell population called neoblasts, which include adult pluripotent stem cells. Thus, it is an excellent model organism for stem cell and regeneration research. Recently, many single-cell RNA sequencing (scRNA-seq) databases of several model organisms, including other planarian species, have become publicly available; these are powerful and useful resources to search for gene expression in various tissues and cells. However, the only scRNA-seq dataset for D. japonica has been limited by the number of genes detected. Herein, we collected D. japonica cells, and conducted an scRNA-seq analysis. A novel, automatic, iterative cell clustering strategy produced a dataset of 3,404 cells, which could be classified into 63 cell types based on gene expression profiles. We introduced two examples for utilizing the scRNA-seq dataset in this study using D. japonica. First, the dataset provided results consistent with previous studies as well as novel functionally relevant insights, that is, the expression of DjMTA and DjP2X-A genes in neoblasts that give rise to differentiated cells. Second, we conducted an integrative analysis of the scRNA-seq dataset and time-course bulk RNA-seq of irradiated animals, demonstrating that the dataset can help interpret differentially expressed genes captured via bulk RNA-seq. Using the R package "Seurat" and GSE223927, researchers can easily access and utilize this dataset.

Dissimilarity-based indicator graph learning for clustering

Although learning a similarity graph with a 0–1 value is regarded as an important issue for clustering tasks, this topic has been scarcely reported in the literature since its NP-hard. Along these lines, this issue is addressed in the work by introducing a novel concept. Concretely, the indicator graph clustering (IGC) method was proposed. Based on the dissimilarity and coordinate descent method, IGC is able to learn an indicator graph with a 0–1 value, which is just the similarity graph with a 0–1 value. Meanwhile, the clustering result can be directly obtained by the indicator graph. From the experimental results on several datasets, it was verified that IGC performs faster than k-means, especially on high-dimensional datasets. Furthermore, IGC was extended to a multi-view version, named indicator graph clustering for multi-view (IGCMV). Our model fused the clustering indicator matrices from different views, which can be obtained by IGC, into the final clustering indicator matrix by a projection matrix. To the best of our knowledge, this work is the first attempt to learn a consensus indicator graph from indicator graphs of all views. In addition different from most of the existing multi-view clustering algorithms, the proposed IGCMV enables the production of the final clustering result, as well as each view’s clustering result simultaneously. This is helpful for measuring the quality of each view. IGCMV can be efficiently optimized by an alternative iteration strategy. Extensive experiments were also carried out on several benchmark multi-view datasets demonstrating the effectiveness of the developed algorithm to the state-of-the-art multi-view clustering algorithms.

J-LOUCR: Joint learned optimized undersampling and constrained reconstruction for accelerated MRI by reference-driven deep image prior

Undersampled magnetic resonance (MR) image reconstruction performance is related to two fundamental problems: where to sample and how to reconstruct the image. Previously developed deep learning-based magnetic resonance imaging (MRI) methods, including undersampled mask optimization and reconstruction, require large-scale, high-quality, and patient-based datasets for training, which presents a challenge in practical clinical scenarios. To reduce the dependence on datasets and perform undersampling and reconstruction simultaneously, we propose a novel joint learned optimized undersampling and constrained reconstruction method for accelerated MRI by a reference-driven deep image prior (J-LOUCR). Inspired by the deep image prior (DIP), J-LOUCR involves no pretraining procedure, resulting in no dependence on large-scale training datasets. The implementation of J-LOUCR includes learned optimization of the undersampled mask and constrained reconstruction with the reference-driven DIP. In the first step, based on the traditional DIP, we introduce a reference image as the network input and propose a joint optimization problem of the undersampled mask and network parameters. Then, the joint optimization problem is deconstructed into subproblems and solved by using the alternative iteration strategy to obtain the optimal undersampled mask guided by the reference image. In the second step, for the optimal undersampled mask, we consider and recommend that the structural similarity between the reference and target MR images be used as prior information. Under the dual drivers of the reference image as the network input and prior constraint, J-LOUCR optimizes the network parameters and then completes reconstruction. Experiments on invivo MR scans demonstrate that J-LOUCR can achieve more accurate reconstruction with more details and features.

Hybrid ant colony and intelligent water drop algorithm for route planning of unmanned aerial vehicles

Route planning is a crucial element in unmanned aerial vehicle (UAV) systems, particularly in autonomous UAV technology. In the past decades, various algorithms have been proposed for UAV route planning. However, they still have defects, such as stagnation and slow search rates. In this study, a novel hybrid algorithm which integrates ant colony optimization (ACO) and intelligent water drop (IWD) is proposed for UAV route planning. First, the advantages of the IWD and ACO algorithms are combined in an iterative strategy, to ensure mutual cooperation via exchange of information for route optimization. Initially, the water drops optimize the soil and pheromones within the environment simultaneously to generate good approximate solutions and an initial pheromone distribution for the ant colony. Based thereupon, the ant colony roams the solution space to further optimize the routes, thus combining the exploration potential and exploitability of two types of agents. Additionally, a novel node selection strategy is proposed to guide the agents’ route planning along a reasonable direction. Compared with state-of-the-art algorithms, the convergence accuracy, success rate, and stability of the proposed algorithm exhibited significant improvements of approximately 8.25%, 4.20%, and 66.20%, respectively.

Maximal Margin Support Vector Machine for Feature Representation and Classification.

High-dimensional small sample size data, which may lead to singularity in computation, are becoming increasingly common in the field of pattern recognition. Moreover, it is still an open problem how to extract the most suitable low-dimensional features for the support vector machine (SVM) and simultaneously avoid singularity so as to enhance the SVM's performance. To address these problems, this article designs a novel framework that integrates the discriminative feature extraction and sparse feature selection into the support vector framework to make full use of the classifiers' characteristics to find the optimal/maximal classification margin. As such, the extracted low-dimensional features from high-dimensional data are more suitable for SVM to obtain good performance. Thus, a novel algorithm, called the maximal margin SVM (MSVM), is proposed to achieve this goal. An alternatively iterative learning strategy is adopted in MSVM to learn the optimal discriminative sparse subspace and the corresponding support vectors. The mechanism and the essence of the designed MSVM are revealed. The computational complexity and convergence are also analyzed and validated. Experimental results on some well-known databases (including breastmnist, pneumoniamnist, colon-cancer, etc.) show the great potential of MSVM against classical discriminant analysis methods and SVM-related methods, and the codes can be available on https://www.scholat.com/laizhihui.