Explore the vast range of titles available.

Quality control processes with automation ensure that customers receive defect-free products that meet their needs. However, the performance of real-world surface defect detection is often severely hindered by the scarcity of data. Recently, few-shot learning has been widely proposed as a solution to the data sufficiency problem by leveraging a limited number of base class samples. However, achieving discriminative and generalization capabilities with few samples remains a challenging task in various surface defect detection scenarios. In this paper, we propose a sparse cross-transformer network (SCTN) for surface defect detection. Specifically, we introduce a residual layer module to enhance the network’s ability to retain crucial information. Next, we propose a sparse layer module within the cross-transformer to increase computational efficiency. Finally, we incorporate a squeeze-and-excitation network into the cross-transformer to enhance the attention mechanism between local patches outputted by the transformer encoder. To verify the effectiveness of our proposed method, we conducted extensive experiments on the cylinder liner defect dataset, the NEU steel surface defect dataset, and the PKU-Market-PCB dataset, achieving the best mean average precision of 62.73%, 85.29%, and 88.7%, respectively. The experimental results demonstrate that our proposed method achieves significant improvements compared to state-of-the-art algorithms. Additionally, the results indicate that SCTN enhances the network’s discriminative ability and effectively improves generalization across various surface defect detection tasks.

In the present study, structural model test using rectangular prismatic granite specimen 200 mm × 200 mm × 200 mm with a horizontal central circular hole of 78 mm diameter was conducted to investigate a rockburst process of borehole. Strain measurement system and high-speed camera were used to capture the rock responses during rockburst. Acoustic emission (AE) system was adopted to monitor the associated AE signals during the rockburst process, and to locate the positions of micro-cracks and subsequently quantitatively investigate the cracking mechanisms. In addition, scanning electron microscope (SEM) was also used to identify the micro-cracks of fragments. The experimental results indicate that rockburst process is characterized by significant spatial distribution and structural responses. As the circumferential stress of surrounding rocks increase, there are some local rockbursts interlacing at different regions of compressive stress concentration along the opening axis direction before overall rockburst. These local rockbursts continued to develop and coalesce, eventually forming overall rockburst. A local rockburst in the present test can be composed of several bursts and persist for a longer period of time than that in true-triaxial tests using rectangular prismatic specimen. Hoop effect, stress gradient around the opening, and last V-shaped bands were accurately simulated. According to AE analysis results, quiet period characterized by few AE hits with high amplitude and a sharp increase in AE energy can be used as an early warning signal for overall rockburst. The time and position of rockburst are related to the spatiotemporal distribution of AE event density, which can be used as a potential indicator for rockburst prediction. During the rockburst process, tensile cracks occupied most of the total micro-cracks, and tensile splitting dominated the failure process. Shear cracks due to tensile cracks interaction initiated at 67% of the spalling strength (tangential stress for spalling failure at the opening boundary) of the borehole.

In the field of artificial intelligence, a large number of promising tools, such as condition-based maintenance, are available for large internal combustion engines. The cylinder liner, which is a key engine component, is subject to defects due to the manufacturing process. In addition, the cylinder liner straightforwardly affects the usage and safety of the internal combustion engine. Currently, the detection of cylinder liner quality mainly depends on manual human detection. However, this type of detection is destructive, time-consuming, and expensive. In this paper, a new cylinder liner defect database is proposed. The goal of this research is to develop a nondestructive yet reliable method for quantifying the surface condition of the cylinder liner. For this purpose, we propose a transformer method with a block division and mask mechanism on our newly collected cylinder liner defect database to automatically detect defects. Specifically, we first use a local defect dataset to train the transformer network. With a hierarchical-level architecture and attention mechanism, multi-level and discriminative feature are obtained. Then, we combine the transformer network with the block division method to detect defects in 64 local regions, and merge their results for the high-resolution image. The block division method can be used to resolve the difficulty of the in detecting the small defect. Finally, we design a mask to suppress the influence of noise. All methods allow us to achieve higher accuracy results than state-of-the-art algorithms. Additionally, we show the baseline results on the new database.

This study sought to develop a radiomics model capable of predicting axillary lymph node metastasis (ALNM) in patients with invasive breast cancer (IBC) based on dual-sequence magnetic resonance imaging(MRI) of diffusion-weighted imaging (DWI) and dynamic contrast enhancement (DCE) data. The interpretability of the resultant model was probed with the SHAP (Shapley Additive Explanations) method. Established inclusion/exclusion criteria were used to retrospectively compile MRI and matching clinical data from 183 patients with pathologically confirmed IBC from our hospital evaluated between June 2021 and December 2023. All of these patients had undergone plain and enhanced MRI scans prior to treatment. These patients were separated according to their pathological biopsy results into those with ALNM ( n  = 107) and those without ALNM ( n  = 76). These patients were then randomized into training ( n  = 128) and testing ( n  = 55) cohorts at a 7:3 ratio. Optimal radiomics features were selected from the extracted data. The random forest method was used to establish three predictive models (DWI, DCE, and combined DWI + DCE sequence models). Area under the curve (AUC) values for receiver operating characteristic (ROC) curves were utilized to assess model performance. The DeLong test was utilized to compare model predictive efficacy. Model discrimination was assessed based on the integrated discrimination improvement (IDI) method. Decision curves revealed net clinical benefits for each of these models. The SHAP method was used to achieve the best model interpretability. Clinicopathological characteristics (age, menopausal status, molecular subtypes, and estrogen receptor, progesterone receptor, human epidermal growth factor receptor 2, and Ki-67 status) were comparable when comparing the ALNM and non-ALNM groups as well as the training and testing cohorts ( P  > 0.05). AUC values for the DWI, DCE, and combined models in the training cohort were 0.793, 0.774, and 0.864, respectively, with corresponding values of 0.728, 0.760, and 0.859 in the testing cohort. The predictive efficacy of the DWI and combined models was found to differ significantly according to the DeLong test, as did the predictive efficacy of the DCE and combined models in the training groups ( P  < 0.05), while no other significant differences were noted in model performance ( P  > 0.05). IDI results indicated that the combined model offered predictive power levels that were 13.5% ( P  < 0.05) and 10.2% ( P  < 0.05) higher than those for the respective DWI and DCE models. In a decision curve analysis, the combined model offered a net clinical benefit over the DCE model. The combined dual-sequence MRI-based radiomics model constructed herein and the supporting interpretability analyses can aid in the prediction of the ALNM status of IBC patients, helping to guide clinical decision-making in these cases.

Using the piezoelectric effect to harvest energy from surrounding vibrations is a promising alternative solution for powering small electronic devices such as wireless sensors and portable devices. A conventional piezoelectric energy harvester (PEH) can only efficiently collect energy within a small range around the resonance frequency. To realize broadband vibration energy harvesting, the idea of multiple-degrees-of-freedom (DOF) PEH to realize multiple resonant frequencies within a certain range has been recently proposed and some preliminary research has validated its feasibility. Therefore, this paper proposed a multi-DOF wideband PEH based on the frequency interval shortening mechanism to realize five resonance frequencies close enough to each other. The PEH consists of five tip masses, two U-shaped cantilever beams and a straight beam, and tuning of the resonance frequencies is realized by specific parameter design. The electrical characteristics of the PEH are analyzed by simulation and experiment, validating that the PEH can effectively expand the operating bandwidth and collect vibration energy in the low frequency. Experimental results show that the PEH has five low-frequency resonant frequencies, which are 13, 15, 18, 21 and 24 Hz; under the action of 0.5 g acceleration, the maximum output power is 52.2, 49.4, 61.3, 39.2 and 32.1 μW, respectively. In view of the difference between the simulation and the experimental results, this paper conducted an error analysis and revealed that the material parameters and parasitic capacitance are important factors that affect the simulation results. Based on the analysis, the simulation is improved for better agreement with experiments.

Nonsuicidal self-injury (NSSI) is one of the major public health problems endangering adolescents. However, the neural mechanisms of NSSI is still unclear. The purpose of this study was to explore regional brain activity and corresponding brain circuits in drug-naïve adolescents with NSSI using amplitude of low-frequency fluctuations (ALFF) combined with functional connectivity (FC) analysis. Thirty-two drug-naïve adolescents with NSSI (NSSI group) and 29 healthy controls matched for sex, age, and level of education (HC group) were enrolled in this study. ALFF and seed-based FC analyses were used to examine the alterations in regional brain activity and corresponding brain circuits. The correlation between ALFF or FC values of aberrant brain regions and clinical characteristics were detected by Pearson correlation analysis. The NSSI group showed increased ALFF in the left inferior and middle occipital gyri, lingual gyrus, and fusiform gyrus; additionally, decreased ALFF in the right medial cingulate gyrus, left anterior cingulate gyrus, and left medial superior frontal gyrus compared to those in the HC group. With the left inferior occipital gyrus as seed, the NSSI group showed increased FC between the left inferior occipital gyrus and the bilateral superior parietal gyrus, right inferior parietal angular gyrus, right inferior frontal gyrus of the insular region, and left precuneus relative to that the HC group. With the left anterior cingulate gyrus as seed, the NSSI group showed increased FC between the left anterior cingulate gyrus and right dorsolateral superior frontal gyrus. With the left lingual gyrus as seed, the NSSI group showed increased FC between the left lingual gyrus and right middle frontal gyrus, and decreased FC between the left lingual gyrus and the left superior temporal gyrus, right supplementary motor area, and left rolandic operculum. With the left fusiform gyrus as seed, the NSSI group showed increased FC between the left fusiform gyrus and left middle and inferior temporal gyrus, and decreased FC between the left fusiform gyrus and the bilateral postcentral gyrus, right precentral gyrus, right lingual gyrus, and left inferior parietal angular gyrus. Moreover, the FC value between the left fusiform gyrus and left inferior temporal gyrus was positively correlated with suicidal ideations score. This study highlights alterations in regional brain activity and corresponding brain circuits in brain regions related to visual and emotional regulation functions in drug-naïve adolescents with NSSI. These findings may facilitate better understand the underlying neural mechanisms of NSSI in adolescents.

Depth estimation based on light field imaging is a new methodology that has succeeded the traditional binocular stereo matching and depth from monocular images. Significant progress has been made in light-field depth estimation. Nevertheless, the balance between computational time and the accuracy of depth estimation is still worth exploring. The geometry in light field imaging is the basis of depth estimation, and the abundant light-field data provides convenience for applying deep learning algorithms. The Epipolar Plane Image (EPI) generated from the light-field data has a line texture containing geometric information. The slope of the line is proportional to the depth of the corresponding object. Considering the light field depth estimation as a spatial density prediction task, we design a convolutional neural network (ESTNet) to estimate the accurate depth quickly. Inspired by the strong image feature extraction ability of convolutional neural networks, especially for texture images, we propose to generate EPI synthetic images from light field data as the input of ESTNet to improve the effect of feature extraction and depth estimation. The architecture of ESTNet is characterized by three input streams, encoding-decoding structure, and skipconnections. The three input streams receive horizontal EPI synthetic image (EPIh), vertical EPI synthetic image (EPIv), and central view image (CV), respectively. EPIh and EPIv contain rich texture and depth cues, while CV provides pixel position association information. ESTNet consists of two stages: encoding and decoding. The encoding stage includes several convolution modules, and correspondingly, the decoding stage embodies some transposed convolution modules. In addition to the forward propagation of the network ESTNet, some skip-connections are added between the convolution module and the corresponding transposed convolution module to fuse the shallow local and deep semantic features. ESTNet is trained on one part of a synthetic light-field dataset and then tested on another part of the synthetic light-field dataset and real light-field dataset. Ablation experiments show that our ESTNet structure is reasonable. Experiments on the synthetic light-field dataset and real light-field dataset show that our ESTNet can balance the accuracy of depth estimation and computational time.

This study aimed to develop an interpretable machine learning model that accurately predicts Ki-67 expression in breast cancer (BC) patients using a combination of dynamic-contrast enhanced magnetic resonance imaging (DCE-MRI) radiomics and clinical-imaging features. A total of 195 BC patients, including 201 lesions, were enrolled retrospectively. These lesions were randomized into training and testing set (7:3). The correlation between clinical-imaging features and Ki-67 expression was analyzed via univariate analysis and binary logistic regression, leading to the development of a Clinical-imaging model. Radiomics features were extracted based on the early and delayed phases of DCE-MRI. These features were screened by Pearson correlation coefficient and recursive feature elimination (RFE). The logistic regression classifier was used to develop the Radiomics model. The clinical imaging and radiomics features were combined to form a Combined model. The Shapley Additive Explanation (SHAP) algorithm was employed to explain the optimal model, and the AUC was used to assess the model’s performance. Ki-67 expression was markedly different from the internal enhancement pattern and necrosis among the imaging features. Compared to the Clinical-imaging model (AUC = 0.682), the AUCs of the Radiomics and the Combined models in the training set were 0.797 and 0.821, respectively. Clinical-imaging, Radiomics, and Combined models had AUCs of 0.666, 0.796, and 0.802 in the test set. Based on the IDI results, the combined model outperformed the Clinical-imaging and Radiomics models in the training set by 11.8% and 2.1%, respectively. They increased by 11% and 1.74% in the test set. SHAP analysis showed that ph2-original-shape-surface volume ratio was the most important feature of the model. Based on clinical-imaging features and DCE-MRI radiomics, the interpretable machine learning model can accurately predict the expression of Ki-67 in BC. Combining the SHAP algorithm with the model improves its interpretability, which may assist clinicians in formulating more accurate treatment strategies.

Long-term, automated fish detection provides invaluable data for deep-sea aquaculture, which is crucial for safe and efficient seawater aquafarming. In this paper, we used an infrared camera installed on a deep-sea truss-structure net cage to collect fish images, which were subsequently labeled to establish a fish dataset. Comparison experiments with our dataset based on Faster R-CNN as the basic objection detection framework were conducted to explore how different backbone networks and network improvement modules influenced fish detection performances. Furthermore, we also experimented with the effects of different learning rates, feature extraction layers, and data augmentation strategies. Our results showed that Faster R-CNN with the EfficientNetB0 backbone and FPN module was the most competitive fish detection network for our dataset, since it took a significantly shorter detection time while maintaining a high AP50 value of 0.85, compared to the best AP50 value of 0.86 being achieved by the combination of VGG16 with all improvement modules plus data augmentation. Overall, this work has verified the effectiveness of deep learning-based object detection methods and provided insights into subsequent network improvements.

Groundwater remediation design is crucial to contemporary water resources management, which is prone to massive computational costs due to the complexity and nonlinearity of the groundwater system. Traditional surrogate methods that can reduce the computational costs tend to encounter barriers of scalability and accuracy when the input–output relationship is highly nonlinear or high-dimensional. To tackle these problems, we herein propose a novel simulation–optimization method that embeds the deep learning deep belief network (DBN) into the particle swarm optimization (PSO) algorithm for groundwater remediation design. Firstly, a numerical simulation model based on MODFLOW and MT3DMS is established to describe the impact on the pollution environmental fate of various implementations of the remediation strategy. The input dataset to train DBN is comprised of various remediation strategies that evolve automatically in the PSO iterations, and the corresponding output dataset constituted of contaminant concentration at observation wells is garnered by executing the simulation model. In the optimization process, the DBN is retrained in an adaptive pattern to enhance prediction accuracy, selectively substituting for the original simulation model to alleviate the computational burden. Additionally, the PSO algorithm undergoes discretization and collision averting within each individual to adapt to the specific remediation task. The results reveal that the proposed method manifests satisfactory convergence behaviour and accuracy, capable of unburdening a considerable proportion (68.8%) of the time consumption for optimal groundwater remediation design.