Maximum Pooling Research Articles

ENHANCING CNN PERFORMANCE WITH RIGHT-NEIGHBOR DEVIATION (RND) POOLING: A FOCUS ON MULTI-CLASS BRAIN TUMOR CLASSIFICATION IN MRI IMAGES

In recent years, the field of image processing and computer vision has been deeply transformed by the advent of Convolutional Neural Network (CNN). These deep learning models, inspired by biological processes, have demonstrated remarkable success in tasks such as image classification, object detection, and semantic segmentation. A critical component of this success is feature extraction, where the pooling layer within CNN serves as a nonlinear down-sampling technique. This layer reduces the spatial size of the convolved feature map through operations like average or maximum pooling, significantly impacting the model’s performance in complex tasks. This research introduces a novel pooling algorithm, Right-Neighbor Deviation (RND), designed to enhance the performance of CNN. The RND pooling method aims to capture detailed spatial features and hierarchical representations more effectively, thereby improving the accuracy and resilience of deep learning models. A dataset of multi-class brain tumor classification, incorporating MRI images from the BR35H dataset, is utilized to explore the potential benefits of RND pooling compared to traditional methods. The effectiveness of the proposed method is evaluated using performance metrics including accuracy, precision, recall, and F1 score, achieving generalized values of 98.86%, 98.84%, 98.80%, and 98.80%, respectively. The results highlight the potential of RND pooling to significantly advance the state of the art in brain tumor classification using CNNs, potentially leading to improved diagnosis and treatment outcomes.

A Novel Detection Transformer Framework for Ship Detection in Synthetic Aperture Radar Imagery Using Advanced Feature Fusion and Polarimetric Techniques

Ship detection in synthetic aperture radar (SAR) imagery faces significant challenges due to the limitations of traditional methods, such as convolutional neural network (CNN) and anchor-based matching approaches, which struggle with accurately detecting smaller targets as well as adapting to varying environmental conditions. These methods, relying on either intensity values or single-target characteristics, often fail to enhance the signal-to-clutter ratio (SCR) and are prone to false detections due to environmental factors. To address these issues, a novel framework is introduced that leverages the detection transformer (DETR) model along with advanced feature fusion techniques to enhance ship detection. This feature enhancement DETR (FEDETR) module manages clutter and improves feature extraction through preprocessing techniques such as filtering, denoising, and applying maximum and median pooling with various kernel sizes. Furthermore, it combines metrics like the line spread function (LSF), peak signal-to-noise ratio (PSNR), and F1 score to predict optimal pooling configurations and thus enhance edge sharpness, image fidelity, and detection accuracy. Complementing this, the weighted feature fusion (WFF) module integrates polarimetric SAR (PolSAR) methods such as Pauli decomposition, coherence matrix analysis, and feature volume and helix scattering (Fvh) components decomposition, along with FEDETR attention maps, to provide detailed radar scattering insights that enhance ship response characterization. Finally, by integrating wave polarization properties, the ability to distinguish and characterize targets is augmented, thereby improving SCR and facilitating the detection of weakly scattered targets in SAR imagery. Overall, this new framework significantly boosts DETR’s performance, offering a robust solution for maritime surveillance and security.

Aircraft engine fault diagnosis based on fused convolutional Transformer

Aircraft engines will face corrosion and erosion problems when working in harsh gas path environments for a long time, and the fault parameter characteristics are not obvious. Therefore, accurate aircraft engine fault diagnosis methods are of great significance to ensure the safe operation of aircraft. In order to improve the prediction accuracy, an aircraft engine fault diagnosis method based on fused convolutional Transformer is proposed. The self-attention mechanism is used to extract useful features and suppress redundant information. The maximum pooling layer (MaxPool) is introduced into the Transformer model to further reduce the model memory consumption and parameter quantity, and alleviate the overfitting phenomenon. The simulation data set of turbofan engines based on GasTurb modeling is used for verification. The results show that compared with the Transformer network and other traditional deep learning models Back Propagation Neural Networks (BP network), Convolutional Neural Networks (CNN) and Recurrent Neural Network (RNN), the accuracy rates are improved by 6.552%、28.117%、13.189%and respectively 10.29%, which proves the effectiveness of the proposed method and can provide a certain reference for aircraft engine fault diagnosis.

Comparison of super-resolution deep learning models for flow imaging

The primary goal of this study is to introduce deep learning (DL) methods as a cost-effective alternative to the computationally intensive Direct Numerical Simulation (DNS) simulations. We show that one can obtain a parametric field from a low-resolution input and map it to a fine grid output, significantly reducing the computational burden. We assess five super-resolution models for up-scaling low-resolution flow data into fine-grid numerical simulations’ output for accuracy and efficiency. The proposed architectures employ convolutional neural networks interconnected in encoder/decoder branches. We investigate these models using turbulent velocity fields inside a suddenly expanded channel characterized by complex features, including turbulence, instabilities, asymmetries, separation, and reattachment. Our results reveal that an encoder/decoder model with residual connections delivers the fastest results, a U-Net-based model with skip connections excels at producing sharper edges in regions prone to blurring, while deeper models incorporating maximum and average pooling layers show superior performance in reconstructing velocity profiles. These findings significantly contribute to our understanding of the potential of deep learning in fluid mechanics. The models presented in this study are trained and validated on standard computer hardware and can be easily adapted to other problems. The findings are promising for discovering and analyzing flow physics, highlighting the potential for DL techniques to improve the accuracy of the available fluid mechanics computational tools.

An improved smoking behavior detection algorithm via incorporating an interference information filtering network

Efficient and accurate identification of smoking behavior in public places is crucial for ensuring public health and safety. However, due to various factors like small target size, complex image background, varying cigarette angles and numerous similar objects, current methods still grapple with challenges such as missed detection and false detection when identifying cigarette targets in smoking behavior. This paper presents an enhanced version of the You Only Look Once version5-small algorithm to address these issues effectively. Firstly, to bolster the model's ability for feature extraction of cigarette targets, the Swin Transformer Block structure is integrated into the backbone network to capture long-range dependencies. Secondly, a novel Hybrid Spatial Pyramid Pooling-Fast with Cross Stage Partial Connection module is built based on the foundation of the Spatial Pyramid Pooling-Fast module by integrating both maximum pooling and average pooling to enhance the fusion ability of the multi-scale feature maps. Thirdly, a novel interference information filtering network is introduced to effectively reduce the impact of noise and confusion caused by similar objects, thus enhancing the performance of the model. According to the experimental results, the accuracy of the improved algorithm on the self-made cigarette target image data set reaches 93.5 %, and the recall reaches 89.1 %.

Research on fault diagnosis algorithm of automobile noise based on deep learning

Abstract This study is devoted to improving the accuracy and efficiency of automobile noise fault diagnosis by using deep learning technology. By constructing a convolutional neural network (CNN) model, we deeply analyze and learn the automobile noise data, aiming at realizing the automatic identification of different types of faults. The experimental results show that the deep learning model shows significant improvement in several performance indicators compared with the traditional methods. Different parameter configurations are used to train the deep learning model. The model structure includes a convolution layer, a maximum pooling layer, and a full connection layer. By extracting the hierarchical features of the data, the model can better adapt to the complex characteristics of noise data. Compared with traditional methods, the CNN model has achieved obvious advantages in accuracy, recall, and F1 score. The research results not only verify the effectiveness of deep learning in automobile noise fault diagnosis but also show the robustness and adaptability of the model under more complex parameter configurations. The conclusion of this study provides empirical support for the application of deep learning technology in the field of automobile noise fault diagnosis.

Current Sensor Fault Detection and Identification for PMSM Drives Using Multichannel Global Maximum Pooling CNN

Current Sensor Fault Detection and Identification for PMSM Drives Using Multichannel Global Maximum Pooling CNN

Hydraulic system fault diagnosis decoupling method based on 2D time-series modeling and self-attention fusion

Hydraulic systems play a pivotal and extensive role in mechanics and energy. However, the performance of intelligent fault diagnosis models for multiple components is often hindered by the complexity, variability, strong hermeticity, intricate structures, and fault concealment in real-world conditions. This study proposes a new approach for hydraulic fault diagnosis that leverages 2D temporal modeling and attention mechanisms for decoupling compound faults and extracting features from multisample rate sensor data. Initially, to address the issue of oversampling in some high-frequency sensors within the dataset, variable frequency data sampling is employed during the data preprocessing stage to resample redundant data. Subsequently, two-dimensional convolution simultaneously captures both the instantaneous and long-term features of the sensor signals for the coupling signals of hydraulic system sensors. Lastly, to address the challenge of feature fusion with multisample rate sensor data, where direct merging of features through maximum or average pooling might dilute crucial information, a feature fusion and decoupling method based on a probabilistic sparse self-attention mechanism is designed, avoiding the issue of long-tail distribution in multisample rate sensor data. Experimental validation showed that the proposed model can effectively utilize samples to achieve accurate fault decoupling and classification for different components, achieving a diagnostic accuracy exceeding 97% and demonstrating robust performance in hydraulic system fault diagnosis under noise conditions.

A Divide-and-Rule Combined Learning Method for Truly Multivariate Time Series Prediction

Multivariate time series prediction is a significant research area that aims to forecast future values based on past observations. Deep learning models with attention mechanisms have shown good predictive performance by emphasizing optimal-related sequences in the target series. However, these models ignore mutation information of nontarget sequences and the long short-term dependencies. To this end, a divide-and-rule combined learning method is proposed to address these limitations, which uses differentiated feature extractors to process different implicit features. First, we design a spatial and temporal information extractor to extract the time-dimensional feature information in the separation stage. Then, a multivariate mutation information extractor is constructed by convolution and maximum pooling layer to capture mutation information of nontarget sequences. Subsequently, the decoder component of the encoder-decoder model extracts long short-term dependencies while preserving the information of the target sequence to be predicted. Finally, in the cooperation stage, a feature fusion method based on a point attention mechanism is proposed, which can assign individual weights to each feature point and enhance the ability to focus on local areas. Experimental results on five real datasets in different domains show that the proposed method has better predictive performance compared to other baseline models.

An Improved U-Net Infrared Small Target Detection Algorithm Based on Multi-Scale Feature Decomposition and Fusion and Attention Mechanism.

Infrared small target detection technology plays a crucial role in various fields such as military reconnaissance, power patrol, medical diagnosis, and security. The advancement of deep learning has led to the success of convolutional neural networks in target segmentation. However, due to challenges like small target scales, weak signals, and strong background interference in infrared images, convolutional neural networks often face issues like leakage and misdetection in small target segmentation tasks. To address this, an enhanced U-Net method called MST-UNet is proposed, the method combines multi-scale feature decomposition and fusion and attention mechanisms. The method involves using Haar wavelet transform instead of maximum pooling for downsampling in the encoder to minimize feature loss and enhance feature utilization. Additionally, a multi-scale residual unit is introduced to extract contextual information at different scales, improving sensory field and feature expression. The inclusion of a triple attention mechanism in the encoder structure further enhances multidimensional information utilization and feature recovery by the decoder. Experimental analysis on the NUDT-SIRST dataset demonstrates that the proposed method significantly improves target contour accuracy and segmentation precision, achieving IoU and nIoU values of 80.09% and 80.19%, respectively.

Fetal electrocardiogram signal extraction based on multi-scale residual shrinkage U-Net

In the extraction of fetal electrocardiogram (ECG) signal, due to the unicity of the scale of the U-Net same-level convolution encoder, the size and shape difference of the ECG characteristic wave between mother and fetus are ignored, and the time information of ECG signals is not used in the threshold learning process of the encoder's residual shrinkage module. In this paper, a method of extracting fetal ECG signal based on multi-scale residual shrinkage U-Net model is proposed. First, the Inception and time domain attention were introduced into the residual shrinkage module to enhance the multi-scale feature extraction ability of the same level convolution encoder and the utilization of the time domain information of fetal ECG signal. In order to maintain more local details of ECG waveform, the maximum pooling in U-Net was replaced by Softpool. Finally, the decoder composed of the residual module and up-sampling gradually generated fetal ECG signals. In this paper, clinical ECG signals were used for experiments. The final results showed that compared with other fetal ECG extraction algorithms, the method proposed in this paper could extract clearer fetal ECG signals. The sensitivity, positive predictive value, and F1 scores in the 2013 competition data set reached 93.33%, 99.36%, and 96.09%, respectively, indicating that this method can effectively extract fetal ECG signals and has certain application values for perinatal fetal health monitoring.

Image reconstruction for electrostatic tomography with deep convolutional neural network

Electrostatic tomography (EST) has the characteristics of being radiation-free and high real-time imaging performance. However, due to its passive measurement method, EST only provides limited input data equal to the number of electrode plates, which results in more serious ill-posed inverse problem. As a result, the imaging accuracy of traditional algorithms is low. To address this problem, an improved deep learning model is proposed for image reconstruction of EST, which is based on a convolutional neural network (CNN). The proposed network is built upon the LeNet-5 deep network model, which incorporates two convolution layers to extract data features and two maximum pooling layers to perform dimension reduction. Furthermore, the LeakyReLU activation function is employed to solve the gradient vanishing problem and accelerate the training process by introducing the non-zero gradients. By constructing an EST model, 8 representative distributions were simulated and 15,930 samples were generated. After 250 iterations, more accurate reconstructed images were generated. To verify the noise immunity of the network model, Gaussian white noises of 20 dB, 30 dB, and 50 dB were added to the test set, and the proposed model could still accurately reconstruct the positions and shapes of the samples. The network is further compared with the traditional algorithms, which reconstructs better results. In addition, the reconstruction results by the proposed model are still relatively accurate for several simulated random samples of different number, shapes and positions. Finally, a 16-electrode sensor array experimental system is designed. The experimental results show that the improved CNN algorithm has a better imaging effect than the traditional algorithms and the CNN algorithm, which is consistent with the simulation results, proving that the model has good generalization ability.

Enhancing museum experience through deep learning and multimedia technology

Amidst the swift progression of artificial intelligence (AI) technology, the museum sector has witnessed a notable inclination towards its adoption. This manuscript endeavours to amplify the interactive milieu of contemporary museum patrons by amalgamating a deep learning algorithm with multimedia technology. The crux of our investigation is the exploration of an adaptive convolutional neural network (CNN) to enrich the interactive engagement of museum visitors. Initially, we leverage the adaptive CNN for the image recognition chore pertaining to museum artifacts and exhibits, thereby facilitating automatic recognition and categorization. Furthermore, to surmount the constraints of conventional pooling algorithms in image feature extraction, we suggest an adaptive pooling algorithm, grounded in the maximum pooling algorithm paradigm. Subsequently, multimedia algorithms are amalgamated into the interactive apparatus, enabling visitors to immerse in exhibits and avail more profound information and experiences. Through juxtaposition with traditional image processing algorithms, the efficacy of our proposed algorithm within a museum ambiance is assessed. Experimental outcomes evince that our algorithm attains superior accuracy and robustness in artifact identification and classification endeavours. In comparison to alternative algorithms, our methodology furnishes more precise and comprehensive displays and interpretations, accurately discerning and categorizing a myriad of exhibit types. This research unveils innovative notions for the digital metamorphosis and advancement of modern museums. Through the incorporation of avant-garde deep learning algorithms and multimedia technologies, the museum visitor experience is elevated, proffering more enthralling and interactive displays. The elucidations of this manuscript hold substantial merit for the continual evolution and innovation within the museum industry.

Convolutional Neural Network to Classify Infrared Thermal Images of Fractured Wrists in Pediatrics.

Convolutional neural network (CNN) models were devised and evaluated to classify infrared thermal (IRT) images of pediatric wrist fractures. The images were recorded from 19 participants with a wrist fracture and 21 without a fracture (sprain). The injury diagnosis was by X-ray radiography. For each participant, 299 IRT images of their wrists were recorded. These generated 11,960 images (40 participants × 299 images). For each image, the wrist region of interest (ROI) was selected and fast Fourier transformed (FFT) to obtain a magnitude frequency spectrum. The spectrum was resized to 100 × 100 pixels from its center as this region represented the main frequency components. Image augmentations of rotation, translation and shearing were applied to the 11,960 magnitude frequency spectra to assist with the CNN generalization during training. The CNN had 34 layers associated with convolution, batch normalization, rectified linear unit, maximum pooling and SoftMax and classification. The ratio of images for the training and test was 70:30, respectively. The effects of augmentation and dropout on CNN performance were explored. Wrist fracture identification sensitivity and accuracy of 88% and 76%, respectively, were achieved. The CNN model was able to identify wrist fractures; however, a larger sample size would improve accuracy.

An attention mechanism module with spatial perception and channel information interaction

In the field of deep learning, the attention mechanism, as a technology that mimics human perception and attention processes, has made remarkable achievements. The current methods combine a channel attention mechanism and a spatial attention mechanism in a parallel or cascaded manner to enhance the model representational competence, but they do not fully consider the interaction between spatial and channel information. This paper proposes a method in which a space embedded channel module and a channel embedded space module are cascaded to enhance the model’s representational competence. First, in the space embedded channel module, to enhance the representational competence of the region of interest in different spatial dimensions, the input tensor is split into horizontal and vertical branches according to spatial dimensions to alleviate the loss of position information when performing 2D pooling. To smoothly process the features and highlight the local features, four branches are obtained through global maximum and average pooling, and the features are aggregated by different pooling methods to obtain two feature tensors with different pooling methods. To enable the output horizontal and vertical feature tensors to focus on different pooling features simultaneously, the two feature tensors are segmented and dimensionally transposed according to spatial dimensions, and the features are later aggregated along the spatial direction. Then, in the channel embedded space module, for the problem of no cross-channel connection between groups in grouped convolution and for which the parameters are large, this paper uses adaptive grouped banded matrices. Based on the banded matrices utilizing the mapping relationship that exists between the number of channels and the size of the convolution kernels, the convolution kernel size is adaptively computed to achieve adaptive cross-channel interaction, enhancing the correlation between the channel dimensions while ensuring that the spatial dimensions remain unchanged. Finally, the output horizontal and vertical weights are used as attention weights. In the experiment, the attention mechanism module proposed in this paper is embedded into the MobileNetV2 and ResNet networks at different depths, and extensive experiments are conducted on the CIFAR-10, CIFAR-100 and STL-10 datasets. The results show that the method in this paper captures and utilizes the features of the input data more effectively than the other methods, significantly improving the classification accuracy. Despite the introduction of an additional computational burden (0.5 M), however, the overall performance of the model still achieves the best results when the computational overhead is comprehensively considered.

An Efficient Homomorphic Argmax Approximation for Privacy-Preserving Neural Networks

Privacy-preserving neural networks offer a promising solution to train and predict without user privacy leakage, and fully homomorphic encryption (FHE) stands out as one of the key technologies, as it enables homomorphic operations over encrypted data. However, only addition and multiplication homomorphisms are supported by FHE, and thus, it faces huge challenges when implementing non-linear functions with ciphertext inputs. Among the non-linear functions in neural networks, one may refer to the activation function, the argmax function, and maximum pooling. Inspired by using a composition of low-degree minimax polynomials to approximate sign and argmax functions, this study focused on optimizing the homomorphic argmax approximation, where argmax is a mathematical operation that identifies the index of the maximum value within a given set of values. For the method that uses compositions of low-degree minimax polynomials to approximate argmax, in order to further reduce approximation errors and improve computational efficiency, we propose an improved homomorphic argmax approximation algorithm that includes rotation accumulation, tree-structured comparison, normalization, and finalization phases. And then, the proposed homomorphic argmax algorithm was integrated into a neural network structure. Comparative experiments indicate that the network with our proposed argmax algorithm achieved a slight increase in accuracy while significantly reducing the inference latency by 58%, as the homomorphic sign and rotation operations were rapidly reduced.

Intelligent algorithm based on deep learning to predict the dosage for anesthesia: A study on prediction of drug efficacy based on deep learning.

Anesthetic drugs play a vital role during surgery, however, due to individual differences and complex physiological mechanisms, the prediction of anesthetic drug dosage has always been a challenging problem. In this study, we propose a model for predicting the dosage of anesthetic drugs based on deep learning to help anesthesiologists better control their dosage during surgical procedures. We design a model based on the artificial neural network to predict the dosage of preoperative anesthetic, and use the SELU activation function and the loss function for weighted regularization to solve the problem of unbalanced sample. Moreover, we design a CNN-based model for the prior extraction of intraoperative features by using a 7 × 1 convolution kernel to enhance the receptive field, and combine maximum pooling and average pooling to extract key features while eliminating noise. A predictive model based on the LSTM network is designed to predict the intraoperative dosage of the anesthetic, and the bidirectional propagation-based LSTM network is used to improve the ability to learn the trend of changes in the physiological states of the patient during surgery. An attention module is added before the connection layer to appropriately attend to areas containing prominent features. The results of experiments showed that the proposed method reduced values of the MAPE to 15.83% and 12.25% compared with the traditional method in predictions of the preoperative and intraoperative doses of the anesthetic, respectively, and increased the values of to 0.887 and 0.915, respectively. The intelligent anesthesia prediction algorithm designed in this study can effectively predict the dosage of anesthetic drugs needed by patients, assist clinical judgment of anesthetic drug dose, and assist the anesthesiologists to ensure the smooth progress of the operation.

A bearing fault diagnosis method based on M-SSCNN and M-LR attention mechanism

Bearing fault diagnosis is vital for mechanical maintenance and fault prediction. It ensures equipment safety, extends lifespan, reduces maintenance costs, and improves production efficiency. Nevertheless, it should be acknowledged that some existing diagnostic methods have achieved high accuracy rates in certain scenarios. However, the challenge lies in their limited generalization capabilities, which can lead to reduced accuracy when applied to diverse or unseen conditions. In this study, we proposed a new bearing fault diagnosis method to address the issue of low accuracy caused by the inadequate generalization of models in the process of rolling bearing fault diagnosis. The method is based on a multi-scale sliding convolution neural network and multi-level residual attention mechanism, the model exhibits high accuracy, strong generalization capability, and lightweight structure. Firstly, the time domain signal of the bearing vibration is converted into a two-dimensional time–frequency map, and the image is pixel-segmented using superpixel segmentation techniques. Next, a multi-scale parallel convolution approach is used to extract features to improve the adaptability and robustness of the model to objects of different sizes and scales. Sliding convolution is used instead of pooling to avoid the problem of feature loss caused by maximum pooling and average pooling. A multi-level attention mechanism is then introduced for all stacked channels to focus on the more important and critical information of the module, and residual connections are added to prevent degradation of the network performance. Finally, the proposed method is passed through the fully connected layer for classification using the Softmax classifier. Experimental verification using public datasets and experimental data of our research group shows that the proposed method has better performance than the existing diagnostic methods and diagnostic models. The proposed method offers an advanced and innovative solution in the domain of bearing fault diagnosis.

Research on inter-shaft bearing fault diagnosis method based on STFT-CNN modeling

Inter-shaft bearings are an essential component of aircraft engines, and their operational status determines the safety of aircraft engine operation. Therefore, to improve the accuracy of fault type prediction and enrich the feature information in vibration signals of aircraft engine inter-shaft bearings, this paper proposes an STFT-CNN model based on the AlexNet architecture, extending its application to the research of aircraft engine inter-shaft bearing fault diagnosis. This approach addresses the common reliance on personnel experience for fault type diagnosis in traditional aircraft engine inter-shaft bearing fault diagnosis. Firstly, real vibration fault signals from inter-shaft bearings are collected through experiments to enrich feature information in non-stationary signals using STFT time-frequency methods. Secondly, utilizing the high interpretability of the STFT-CNN model, fault feature data from inter-shaft bearings under various operating conditions are extracted to refine our understanding of fault feature information. Finally, leveraging the robustness of the STFT-CNN model, fault types are classified and predicted. The training process involves comparative analysis using different pooling algorithms, time-frequency analysis methods, and various deep learning network models. The results demonstrate that the STFT-CNN model, employing the maximum pooling algorithm, outperforms other models in predicting inter-shaft bearing faults, achieving an average fault prediction accuracy of 98.8% .

MACN: A cascade defect detection for complex background based on mixture attention mechanism

AbstractDefect detection in complex background is a critical issue. To address this issue, this paper proposes the mixture attention mechanism cascade network, in which the new channel attention network is linked with the spatial attention network to create an effective mixed attention network that takes advantage of their respective advantages, adaptively suppresses background noise, and highlights defect features. To ensure the efficiency and effectiveness of effective mixed attention network, the new channel attention network splices the output features of the global average pooling layer and the global maximum pooling layer and then sends the spliced features into a shared network, which is a one‐dimensional convolutional network, and uses cross‐channel interaction for fusion. Furthermore, in order to provide more discriminative feature representation, the authors extract the intermediate features of the region proposal network and input them into effective mixed attention network. Finally, the cascade head is used to refine the predicted bounding box to achieve high‐quality defect location. To demonstrate the superiority and usefulness of this method, it is compared to the latest method using widely used PCB and NEU data sets. A large number of trials demonstrate that this strategy outperforms other methods for detecting defects in complicated backgrounds.