Complex & Intelligent Systems最新文献

Failure mode and effects analysis (FMEA) is an important risk analysis tool that has been widely used in diverse areas to manage risk factors. However, how to manage the uncertainty in FMEA assessments is still an open issue. In this paper, a novel FMEA model based on the improved pignistic probability transformation function in Dempster–Shafer evidence theory (DST) and grey relational projection method (GRPM) is proposed to improve the accuracy and reliability in risk analysis with FMEA. The basic probability assignment (BPA) function in DST is used to model the assessments of experts with respect to each risk factor. Dempster’s rule of combination is adopted for fusion of assessment information from different experts. The improved pignistic probability function is proposed and used to transform the fusion result of BPA into probability function for getting more accurate decision-making result in risk analysis with FMEA. GRPM is adopted to determine the risk priority order of all the failure modes to overcome the shortcoming in traditional risk priority number in FMEA. Applications in aircraft turbine rotor blades and steel production process are presented to show the rationality and generality of the proposed method.

The security and confidentiality of medical images are of utmost importance due to frequent issues such as leakage, theft, and tampering during transmission and storage, which seriously impact patient privacy. Traditional encryption techniques applied to entire images have proven to be ineffective in guaranteeing timely encryption and preserving the privacy of organ regions separated from the background. In response, this study proposes a specialized and efficient local image encryption algorithm for the medical field. The proposed encryption algorithm focuses on the regions of interest (ROI) within massive medical images. Initially, the Laplacian of Gaussian operator and the outer boundary tracking algorithm are employed to extract the binary image and achieve ROI edge extraction. Subsequently, the image is divided into ROI and ROB (regions outside ROI). The ROI is transformed into a row vector and rearranged using the Lorenz hyperchaotic system. The rearranged sequence is XOR with the random sequence generated by the Henon chaotic map. Next, the encrypted sequence is arranged according to the location of the ROI region and recombined with the unencrypted ROB to obtain the complete encrypted image. Finally, the least significant bit algorithm controlled by the key is used to embed binary image into the encrypted image to ensure lossless decryption of the medical images. Experimental verification demonstrates that the proposed selective encryption algorithm for massive medical images offers relatively ideal security and higher encryption efficiency. This algorithm addresses the privacy concerns and challenges faced in the medical field and contributes to the secure transmission and storage of massive medical images.

For visual estimation of optical flow, which is crucial for various vision analyses, unsupervised learning by view synthesis has emerged as a promising alternative to supervised methods because the ground-truth flow is not readily available in many cases. However, unsupervised learning is likely to be unstable when pixel tracking is lost via occlusion and motion blur, or pixel correspondence is impaired by variations in image content and spatial structure over time. Recognizing that dynamic occlusions and object variations usually exhibit a smooth temporal transition in natural settings, we shifted our focus to model unsupervised learning optical flow from multi-frame sequences of such dynamic scenes. Specifically, we simulated various dynamic scenarios and occlusion phenomena based on Markov property, allowing the model to extract motion laws and thus gain performance in dynamic and occluded areas, which diverges from existing methods without considering temporal dynamics. In addition, we introduced a temporal dynamic model based on a well-designed spatial-temporal dual recurrent block, resulting in a lightweight model structure with fast inference speed. Assuming the temporal smoothness of optical flow, we used the prior motions of adjacent frames to supervise the occluded regions more reliably. Experiments on several optical flow benchmarks demonstrated the effectiveness of our method, as the performance is comparable to several state-of-the-art methods with advantages in memory and computational overhead.

This paper researches the problem of Beyond Visual Range (BVR) air combat maneuver intention recognition. To achieve efficient and accurate intention recognition, an Attention enhanced Tuna Swarm Optimization-Parallel Bidirectional Gated Recurrent Unit network (A-TSO-PBiGRU) is proposed, which constructs a novel Parallel BiGRU (PBiGRU). Firstly, PBiGRU has a parallel network structure, whose proportion of forward and backward network can be adjusted by forward coefficient and backward coefficient. Secondly, to achieve object-oriented adjustment of forward and backward coefficients, the tuna swarm optimization algorithm is introduced and the negative log-likelihood estimation loss function is used as the objective function, it realizes the dynamic combination of sequence guidance and reverse correction. Finally, the attention mechanism is used to obtain more useful information to improve the recognition accuracy. Through offline recognition experiment, it is proved that A-TSO-PBiGRU can effectively improve the convergence speed and recognition accuracy compared with GRU-related networks. Compared with the other six comparison algorithms, maneuver intention recognition accuracy also has significant advantages. In the online recognition experiment, maneuver intention recognition accuracy of A-TSO-PBiGRU is 93.7%, it shows excellent maneuver intention recognition ability.

Social image data refer to the annotated image with tags in social media, in which the tags are always labeled by users. Integrating the visual and textual information of social image can obtain accurate and comprehensive feature and improve clustering performance. However, the heterogeneous gap between tags and images makes it difficult to reasonably organize the social images. In addition, the tags are often sparse and incomplete due to personal preference and cognition differences of users. To solve these problems, we propose a novel knowledge-aware progressive clustering (KAPC) method, which employs human knowledge to guide the cross-modal clustering of social images. Firstly, we design a dual-similarity semantic expansion strategy to complement the sparse tags with human knowledge, which constructs a more complete semantic similarity matrix for tags through knowledge graphs. Secondly, we define an objective function based on information theory to bridge the heterogeneous gap, which align inter-modal cluster distribution to explore the correlation between visual and textual information. Finally, a progressive iteration method is designed to make the two modalities guide each other and obtain better performance of social image clustering. Extensive experiments on four social image datasets verify the effectiveness of the proposed KAPC method.

This article presents a detailed investigation into the Multi-Depot Half-Open Time-Dependent Electric Vehicle Routing Problem (MDHOTDEVRP) within the domain of urban distribution, prompted by the growing urgency to mitigate the environmental repercussions of logistics transportation. The study first surmounts the uncertainty in Electric Vehicle (EV) range arising from the dynamic nature of urban traffic networks by establishing a flexible energy consumption estimation strategy. Subsequently, a Mixed-Integer Programming (MIP) model is formulated, aiming to minimize the total distribution costs associated with EV dispatch, vehicle travel, customer service, and charging operations. Given the unique attributes intrinsic to the model, a Two-Stage Hybrid Ant Colony Algorithm (TSHACA) is developed as an effective solution approach. The algorithm leverages enhanced K-means clustering to assign customers to EVs in the first stage and employs an Improved Ant Colony Algorithm (IACA) for optimizing the distribution within each cluster in the second stage. Extensive simulations conducted on various test scenarios corroborate the economic and environmental benefits derived from the MDHOTDEVRP solution and demonstrate the superior performance of the proposed algorithm. The outcomes highlight TSHACA’s capability to efficiently allocate EVs from different depots, optimize vehicle routes, reduce carbon emissions, and minimize urban logistic expenditures. Consequently, this study contributes significantly to the advancement of sustainable urban logistics transportation, offering valuable insights for practitioners and policy-makers.

This paper proposes a novel algorithm named surrogate ensemble assisted differential evolution with efficient dual differential grouping (SEADECC-EDDG) to deal with large-scale expensive optimization problems (LSEOPs) based on the CC framework. In the decomposition phase, our proposed EDDG inherits the framework of efficient recursive differential grouping (ERDG) and embeds the multiplicative interaction identification technique of Dual DG (DDG), which can detect the additive and multiplicative interactions simultaneously without extra fitness evaluation consumption. Inspired by RDG2 and RDG3, we design the adaptive determination threshold and further decompose relatively large-scale sub-components to alleviate the curse of dimensionality. In the optimization phase, the SEADE is adopted as the basic optimizer, where the global and the local surrogate model are constructed by generalized regression neural network (GRNN) with all historical samples and Gaussian process regression (GPR) with recent samples. Expected improvement (EI) infill sampling criterion cooperated with random search is employed to search elite solutions in the surrogate model. To evaluate the performance of our proposal, we implement comprehensive experiments on CEC2013 benchmark functions compared with state-of-the-art decomposition techniques. Experimental and statistical results show that our proposed EDDG is competitive with these advanced decomposition techniques, and the introduction of SEADE can accelerate the convergence of optimization significantly.

The design of binder proteins for specific target proteins using deep learning is a challenging task that has a wide range of applications in both designing therapeutic antibodies and creating new drugs. Machine learning-based solutions, as opposed to laboratory design, streamline the design process and enable the design of new proteins that may be required to address new and orphan diseases. Most techniques proposed in the literature necessitate either domain knowledge or some appraisal of the target protein’s 3-D structure. This paper proposes an approach for designing binder proteins based solely on the amino acid sequence of the target protein and without recourse to domain knowledge or structural information. The sequences of the binders are generated with two new transformers, namely the AppendFormer and MergeFormer architectures. Because, in general, there is more than one binder for a given target protein, these transformers employ a binding score and a prior on the sequence of the binder to obtain a unique targeted solution. Our experimental evaluation confirms the strengths of this novel approach. The performance of the models was determined with 5-fold cross-validation and clearly indicates that our architectures lead to highly accurate results. In addition, scores of up to 0.98 were achieved in terms of Needleman-Wunsch and Smith-Waterman similarity metrics, which indicates that our solutions significantly outperform a seq2seq baseline model.

Light guide plate (LGP) is a key component of liquid crystal display (LCD) display systems, so its quality directly affects the display effect of LCD. However, LGPs have complex background texture, low contrast, varying defect size and numerous defect types, which makes realizing efficient and accuracy-satisfactory surface defect automatic detection of LGPS still a big challenge. Therefore, combining its optical properties, dot distribution, defect imaging characteristics and detection requirements, a surface defect detection algorithm based on LGP-YOLO for practical industrial applications is proposed in this paper. To enhance the feature extraction ability of the network without dimensionality reduction, expand the effective receptive field and reduce the interference of invalid targets, we built the receptive field module (RFM) by combining the effective channel attention network (ECA-Net) and reviewing large kernel design in CNNs (RepLKNet). For the purpose of optimizing the performance of the network in downstream tasks, enhance the network's expression ability and improve the network’s ability of detecting multi-scale targets, we construct the small detection module (SDM) by combining space-to-depth non-strided convolution (SPDConv) and omini-dimensional dynamic convolution (ODConv). Finally, an LGP defect dataset is constructed using a set of images collected from industrial sites, and a multi-round experiment is carried out to test the proposed method on the LGP detect dataset. The experimental results show that the proposed LGP-YOLO network can achieve high performance, with mAP and F1-score reaching 99.08% and 97.45% respectively, and inference speed reaching 81.15 FPS. This demonstrates that LGP-YOLO can strike a good balance between detection accuracy and inference speed, capable of meeting the requirements of high-precision and high-efficiency LGP defect detection in LGP manufacturing factories.

During pineapple cultivation, detecting and counting the number of pineapple flowers in real time and estimating the yield are essential. Deep learning methods are more efficient in real-time performance than traditional manual detection. However, existing deep learning models are characterized by low detection speeds and cannot be applied in real time on mobile devices. This paper presents a lightweight model in which filter pruning compresses the YOLOv5 network. An adaptive batch normalization layer evaluation mechanism is introduced to the pruning process to evaluate the performance of the subnetwork. With this approach, the network with the best performance can be found quickly after pruning. Then, an efficient channel attention mechanism is added for the pruned network to constitute a new YOLOv5_E network. Our findings demonstrate that the proposed YOLOv5_E network attains an accuracy of 71.7% with a mere 1.7 M parameters, a model size of 3.8 MB, and an impressive running speed of 178 frames per second. Compared to the original YOLOv5, YOLOv5_E shows a 0.9% marginal decrease in accuracy; while, the number of parameters and the model size are reduced by 75.8% and 73.8%, respectively. Moreover, the running speed of YOLOv5_E is nearly twice that of the original. Among the ten networks evaluated, YOLOv5_E boasts the fastest detection speed and ranks second in detection accuracy. Furthermore, YOLOv5_E can be integrated with StrongSORT for real-time detection and counting on mobile devices. We validated this on the NVIDIA Jetson Xavier NX development board, where it achieved an average detection speed of 24 frames per second. The proposed YOLOv5_E network can be effectively used on agricultural equipment such as unmanned aerial vehicles, providing technical support for the detection and counting of crops on mobile devices.