Applied Soft Computing最新文献

{"title":"A quantum entanglement-based optimization method for complex expensive engineering problems","authors":"Fengling Peng,&nbsp;Xing Chen","doi":"10.1016/j.asoc.2025.113019","DOIUrl":"10.1016/j.asoc.2025.113019","url":null,"abstract":"<div><div>Due to the computational costliness and time-consuming nature of complex and expensive engineering (CEE) problems, this paper proposes a genetic algorithm based on quantum entanglement to address these challenges. This method encodes individuals into quantum genes, where each gene bit stores not 0 or 1, but a superposition state of both. By leveraging the uncertainty of the superposition state during the collapse, this method effectively preserves population diversity even with a very small population size. A smaller population size implies fewer calls to time-consuming simulations. Additionally, quantum entangled states are created for parts of an individual's gene, utilizing the characteristic that entangled states instantly affect each other upon collapse, to achieve parallel evolution of parts of the genes in multiple individuals. This parallel evolution significantly increases the search speed of the algorithm, thereby reducing the number of iterations. Fewer iterations also mean fewer calls to simulations. Benchmark function experiments demonstrate that the proposed method is significantly superior to other similar algorithms in a 30D solution space with a population size of 20 and also has certain advantages in a 100D solution space.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"174 ","pages":"Article 113019"},"PeriodicalIF":7.2,"publicationDate":"2025-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143687318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-reference super-resolution reconstruction of remote sensing images based on hierarchical similarity mapping","authors":"Fuzhen Zhu ,&nbsp;Qi Zhang ,&nbsp;Bing Zhu ,&nbsp;Chen Wang","doi":"10.1016/j.asoc.2025.113027","DOIUrl":"10.1016/j.asoc.2025.113027","url":null,"abstract":"<div><div>To make full use of the details from multi-reference images and improve the quality of super-resolution reconstruction of remote sensing images, a multi-reference super-resolution reconstruction of remote sensing images based on hierarchical similarity mapping is proposed. It is very important in both military and civilian fields. Firstly, one low resolution image and three reference images are used as the input of VGG network to extract their feature maps at 4 × , 2 × , and 1 × scales. These feature maps at each scale are respectively blocked and used as a set of inputs in subsequent operations. Specifically, the low resolution features are divided into <span><math><msub><mrow><mi>N</mi></mrow><mrow><mi>i</mi></mrow></msub></math></span>blocks, and each block is further divided into <span><math><msub><mrow><mi>N</mi></mrow><mrow><mi>c</mi></mrow></msub></math></span>sub-feature-blocks. And the <span><math><msub><mrow><mi>N</mi></mrow><mrow><mi>m</mi></mrow></msub></math></span>reference image features are divided into <span><math><msub><mrow><mi>N</mi></mrow><mrow><mi>r</mi></mrow></msub></math></span> sub-feature-blocks. Then the <span><math><msub><mrow><mi>N</mi></mrow><mrow><mi>c</mi></mrow></msub></math></span>low-resolution sub-feature blocks are mapped for similarity with the reference features within the range of all reference sub-feature blocks, individual reference features, and all reference image features. The outputs of each layer are then iteratively mapped with the low-resolution features as inputs for next layers. Thus the final features include information from all the reference images and low-resolution image. Subsequently, an adaptive transfer module with multi-reference features and channel attention is used to match and transfer the information of each reference image, while achieving edge smoothing and noise filtering between different reference features. Finally, the quadruple super-resolution reconstruct result is got from the multi-scale feature fusion module and decoder. Experimental results show that our improvements can reconstruct better super-resolution results with more details for utilizing information of multi-reference images, which is superior to single image super-resolution methods and single reference super-resolution methods.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"174 ","pages":"Article 113027"},"PeriodicalIF":7.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143724733","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A morphological difference and statistically sparse transformer-based deep neural network for medical image segmentation","authors":"Dongxu Cheng ,&nbsp;Zifang Zhou ,&nbsp;Hao Li ,&nbsp;Jingwen Zhang ,&nbsp;Yan Yang","doi":"10.1016/j.asoc.2025.113052","DOIUrl":"10.1016/j.asoc.2025.113052","url":null,"abstract":"<div><div>Medical image segmentation plays a pivotal role in enhancing disease diagnosis and treatment planning. However, existing methods often struggle with the complexity of lesion boundaries and the computational demands of Transformer-based approaches. To address these challenges, we propose a morphological difference and statistically sparse Transformer-based deep neural network for medical image segmentation, termed MD-SSFormer. It comprises two critical modules: the dual branch encoder (DBEncoder) module, and the morphological difference catcher (MDC). To extract abundant information at different aspects, a novel DBEncoder module integrates the capability of the convolutional neural network-based method in capturing local texture and the ability of the Transformer-based method in modeling global information. Compared to the conventional feature extraction methods, DBEncoder achieves comprehensive improvement. Furthermore, the statistics-based sparse Transformer (SSFormer) module develops an innovative statistical analysis and an adaptive patch-dividing strategy to perform attention-computing, which addresses the computational challenges associated with conventional Transformer-based models. Finally, considering the impacts of the blurry and complex boundaries, the MDC module employs the morphological operation and differential information extractor to refine the details, which achieves high-precision boundary understanding. Experimental results on five public datasets demonstrate MD-SSFormer's superior performance, achieving state-of-the-art Dice scores of 83.60 % on ISIC 2017, 79.52 % on Kvasir-SEG, 61.89 % on BUSI, 78.62 % on BraTS21, and 85.85 % on 3DIRCADb, outperforming other methods in accuracy, precision, and computational efficiency respectively.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"174 ","pages":"Article 113052"},"PeriodicalIF":7.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143698029","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Integrating Competitive Framework into Differential Evolution: Comprehensive performance analysis and application in brain tumor detection","authors":"Rui Zhong ,&nbsp;Zhongmin Wang ,&nbsp;Yujun Zhang ,&nbsp;Junbo Jacob Lian ,&nbsp;Jun Yu ,&nbsp;Huiling Chen","doi":"10.1016/j.asoc.2025.112995","DOIUrl":"10.1016/j.asoc.2025.112995","url":null,"abstract":"<div><div>This paper presents an efficient and effective optimizer based on the Success History Adaptive DE (SHADE) named Competitive Framework DE (CFDE). We integrate three tailored strategies into CFDE: (1) the competitive framework to identify and prioritize potential individuals, (2) the novel DE/loser-to-best/loser-to-winner mutation scheme to fully leverage the information from the population and competition to construct high-quality offspring individuals, and (3) the random memory initialization to diversify the search patterns of the individual. We conduct comprehensive numerical experiments on CEC2017, CEC2020, CEC2022, and eight engineering problems against eleven state-of-the-art optimizers to confirm the superiority and competitiveness of CFDE. Moreover, the sensitivity experiments on hyperparameters validate the robustness of CFDE, and the ablation experiments practically prove the independent contribution of integrated components. Furthermore, we propose a hybrid model named DenseNet-CFDE-ELM for brain tumor detection, where DenseNet-169 is employed for feature selection and CFDE-optimized Extreme Learning Machine (ELM) classifies the brain tumors in MRI scans. Experimental results on the brain tumor dataset downloaded from Kaggle confirm that the proposed DenseNet-CFDE-ELM achieves improvements in accuracy with 1.794%, precision with 1.696%, recall with 1.794%, and F1 score with 1.812% against the second-best ResNet-18 model. These results reveal the potential of CFDE in extensive real-world optimization scenarios. The source code of this research can be downloaded from <span><span>https://github.com/RuiZhong961230/CFDE</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"174 ","pages":"Article 112995"},"PeriodicalIF":7.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143724711","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"State-space recurrent neural networks for predictive analytics and latent state estimation","authors":"Ramin Moghaddass,&nbsp;Cheng-Bang Chen","doi":"10.1016/j.asoc.2025.113017","DOIUrl":"10.1016/j.asoc.2025.113017","url":null,"abstract":"<div><div>This paper presents a framework to predict the remaining life (RL) of degrading systems under sensor condition monitoring. By integrating state-space modeling with stochastic recurrent neural networks, our approach efficiently processes condition-monitoring time-series data and models systems’ latent degradation states. We propose a stochastic model that captures dependencies among latent degradation states, sensor outputs, and RL in a causally coherent manner and utilizes stochastic neural networks to navigate the inherent uncertainties of system dynamics. To enhance the interpretability of RL estimation and latent state modeling, we propose interpretable regularization terms. These terms are incorporated into the loss function to optimize both the prediction precision of estimating remaining life and latent states and control the monotonic behavior of their estimates, thereby improving the model’s overall performance and interpretability. Our methodology is validated through numerical experiments and comparison with benchmark models, demonstrating its potential to improve predictive maintenance strategies by effectively estimating the remaining life and monitoring the state of latent degradation over time.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"174 ","pages":"Article 113017"},"PeriodicalIF":7.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143686933","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"USVDD: A one-class outlier detection method with imprecisely- observed data","authors":"Jingyu Liang,&nbsp;Jie Liu","doi":"10.1016/j.asoc.2025.113065","DOIUrl":"10.1016/j.asoc.2025.113065","url":null,"abstract":"<div><div>Detecting outliers is essential to recognizing potential faults in high-value equipment. Since such equipment typically undergoes careful maintenance to prevent breakdowns, outlier detection often involves a one-class classification problem. Current advanced methods frequently face difficulties in managing the complexities of one-class observations, particularly when dealing with uncertain and imprecise monitoring data. This study proposes a novel approach to fault detection, called uncertain support vector data description (USVDD), in contexts involving one-class uncertain data. Drawing on uncertainty theory, USVDD conceptualizes imprecise observations as uncertain variables. By integrating novel hyperparameters, particularly belief degrees <span><math><msub><mrow><mi>α</mi></mrow><mrow><mi>k</mi></mrow></msub></math></span>, the method effectively addresses and quantifies data uncertainty, enhancing its ability to handle complex, uncertain datasets. Testing on 12 real-world datasets highlights that the proposed USVDD method outperforms alternative methods by achieving the highest balanced F1-scores on 10 datasets. Additionally, it demonstrates remarkable efficiency, completing computations on high-dimensional datasets up to 60 times faster than competing methods on the same hardware setup. USVDD excels in managing observational data influenced by aleatory uncertainty, making it a reliable solution for one-class diagnostic modeling in situations with scarce fault samples.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"175 ","pages":"Article 113065"},"PeriodicalIF":7.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143746221","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Arithmetic optimization algorithm with cosine composite chaotic mapping in polar coordinate system for economic load dispatching problems in power systems","authors":"Yi-Xuan Li,&nbsp;Jie-Sheng Wang,&nbsp;Si-Wen Zhang,&nbsp;Shi-Hui Zhang,&nbsp;Xin-Yi Guan,&nbsp;Xin-Ru Ma","doi":"10.1016/j.asoc.2025.113039","DOIUrl":"10.1016/j.asoc.2025.113039","url":null,"abstract":"<div><div>Economic load dispatch (ELD) aims to minimize the total cost of generating electricity while satisfying load demand and different operational constraints. The Arithmetic Optimization Algorithm (AOA) with cosine composite chaotic mapping in polar coordinate system is put forward to solve the ELD problems in the power system with the valve point effect, prohibited operation area, transmission loss and other factors. Firstly, seven polar coordinate system chaotic mappings are proposed to be embedded into the MOP and MOA parameters in the AOA. Secondly, a chaotic system based on the cosine transform is put forward. Then, the proposed cosine transform based chaotic system is combined with polar coordinate system chaotic mapping to form polar coordinate system cosine transform composite chaotic mapping. Eventually, these six polar coordinate system cosine transform composite chaotic mapping is then embedded into the MOA and MOP of the AOA to balance the algorithm's global and local search capabilities, improve the performance of the algorithm and avoid falling into the local optima. The superiority of the improved algorithm is verified by employing 12 benchmark test functions in CEC2022. Then, it is compared with the Coati Optimization Algorithm (COA), Prairie Dog Optimization (PDO), Butterfly Optimization Algorithm (BOA), Reptile Search Algorithm (RSA), Bat Algorithm (BAT) and Osprey Optimization Algorithm (OOA) to verify its convergence. The ELD problem for a total demand of 2500 MW is solved by using the AOA with cosine composite chaotic mapping in polar coordinate system. The experimental results show that the improved AOA outperforms other optimization algorithms on the 12 benchmark functions in CEC2022 and the ELD problems.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"175 ","pages":"Article 113039"},"PeriodicalIF":7.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143738483","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A multi-branch attention coupled convolutional domain adaptation network for bearing intelligent fault recognition under unlabeled sample scenarios","authors":"Maoyou Ye,&nbsp;Xiaoan Yan,&nbsp;Dong Jiang,&nbsp;Ning Chen","doi":"10.1016/j.asoc.2025.113053","DOIUrl":"10.1016/j.asoc.2025.113053","url":null,"abstract":"<div><div>Bearing intelligent fault recognition is important to maintain the healthy and stable operation of mechanical equipment. However, it is difficult to have a consistent distribution of the acquired source and target domain data due to the constantly changing operating state of the equipment. Moreover, the acquisition of sufficient labeled data is constrained by both time and economic costs. Most of the existing recognition methods are difficult to perform effective fault recognition when faced with inconsistent data distribution and unlabeled small sample data. To address these issues, this paper proposes a multi-branch attention coupled convolutional domain adaptation network (MACCDAN) for unsupervised cross-domain fault recognition, which contains three unique parts. A cross-attention coupled module (CACM) is firstly designed between two parallel feature extraction branches to guide the intertwined coupling of the two branch features through a dual synergetic attention mechanism. A global feature aggregation module (GFAM) is further presented to conduct the global information fusion, which integrates the dependencies between different branch features and enhances the perception of key features. Additionally, the maximum-similarity minimum-discrepancy adversarial loss (MSMDAL) is formulated as an optimization objective to reduce the discrepancy between the source and target domain, and promote the learning of domain-invariant and discriminative features. The results of the four performance evaluation metrics (i.e., accuracy, precision, recall and F1 score) of the proposed method are all 1.0000 on two datasets. The F1 score of the proposed method is improved by at least 0.03 compared to other methods.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"174 ","pages":"Article 113053"},"PeriodicalIF":7.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143686929","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Category-level pipe pose and size estimation via geometry-aware adaptive curvature convolution","authors":"Jia Hu,&nbsp;Jianhua Liu,&nbsp;Shaoli Liu,&nbsp;Lifeng Wang","doi":"10.1016/j.asoc.2025.113006","DOIUrl":"10.1016/j.asoc.2025.113006","url":null,"abstract":"<div><div>Pipe pose estimation provides crucial positional information for robots, enhancing assembly efficiency and precision, while its accuracy critically impacts the final product's reliability and quality. To handle unseen pipes, we propose a category-level pipe pose and size estimation network via Normalized Object Coordinate Space (NOCS) representation. Given an RGB image and its corresponding depth map, our network predicts class labels, bounding boxes and instance masks for detection, as well as NOCS maps for pose estimation. Then these predictions are aligned with the depth map to estimate pipe’s pose and size. To better extract complex and variable pipe morphology, geometry-aware adaptive curvature convolution is introduced to dynamically adapt to the slender structure and improve segmentation performance. Facing the lack of pipe pose datasets with enough instances, pose, clutter, occlusion, and illumination variation, we propose a novel domain randomization mixed reality approach to efficiently generate synthetic data, which addresses the limitations of training datasets, making data generation more time- and effort-efficient. Experimental results demonstrate that our Geometry-Aware Adaptive Convolutional Network (GACNet) outperforms other methods and robustly estimates the pose and size of unseen pipes in real-world environments.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"174 ","pages":"Article 113006"},"PeriodicalIF":7.2,"publicationDate":"2025-03-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143686934","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A virtual try-on network with arm region preservation","authors":"Xin Zhang,&nbsp;Jinguang Chen,&nbsp;Lili Ma,&nbsp;Kaibing Zhang","doi":"10.1016/j.asoc.2025.112960","DOIUrl":"10.1016/j.asoc.2025.112960","url":null,"abstract":"<div><div>Virtual try-on methods based on image generation have made some progress. However, the try-on results obtained are unnatural due to the difficulty in generating details such as hands. In order to address this issue, we introduce an arm-region preserving virtual try-on network, AP-VITON. First, we constructed an arm region preservation module to achieve the extraction and preservation of arm preservable regions through a series of image processing operations. Second, in order to enhance the image generation process, we combine the try-on generator module with a lightweight vision transformer, MobileViT, with the objective of improving the network’ s ability to capture global information. Finally, we introduce focal frequency loss to overcome the limitation of conventional methods that use only spatial domain loss. Quantitative and qualitative comparisons on a generic virtual try-on dataset show that our approach produces more realistic try-on results, in particular a 36.6% improvement in the KID metric and a 29.5% improvement in the LPIPS metric.</div></div>","PeriodicalId":50737,"journal":{"name":"Applied Soft Computing","volume":"175 ","pages":"Article 112960"},"PeriodicalIF":7.2,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143738476","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}