Few-shot Learning for Bagel Defect Detection

In this paper, we equip Prototypical Networks (PNs) with a novel dissimilarity measure to enable discriminative feature normalization for few-shot learning. The embedding onto the hypersphere requires no direct normalization and is easy to optimize. Our theoretical analysis shows that the proposed dissimilarity measure, denoted the Squared root of the Euclidean distance and the Norm distance (SEN), forces embedding points to be attracted to its correct prototype, while being repelled from all other prototypes, keeping the norm of all points the same. The resulting SEN PN outperforms the regular PN with a considerable margin, with no additional parameters as well as with negligible computational overhead.

Accurate identification of pathogenic bacteria from images taken under light microscopy is an important role in clinical diagnosis and microbiology. Traditional deep learning methods such as Convolutional Neural Networks (CNNs) have achieved excellent performance but are based on enormous, well-annotated datasets that may not be readily available. Absence of labeled images of bacteria is one of the key challenges in real-world applications, where class imbalance and diversity of samples affect generalization. Recent years have witnessed few-shot learning being a potential paradigm for mitigating data paucity by enabling models to learn well from sparse samples. Among numerous few-shot approaches, prototypical networks have proven useful for image classification with sparse data. This paper overviews the progress in bacterial identification from light microscopy images, reviews existing classical and deep learning methods, and highlights how prototypical networks solve the shortcomings of conventional models through few-shot learning. The review concludes with open problems and future directions, including integrating fewshot learning into web-based, scalable diagnostic platforms for efficient pathogen detection.

Facial skin type analysis is a critical task in several fields, including dermatology, cosmetics, and biometrics, and has been the subject of significant research in recent years. Traditional facial skin type analysis approaches rely on large, labeled datasets, which can be time-consuming and costly to collect. This study proposes a novel few-shot learning (FSL) approach for facial skin type analysis that can accurately classify skin types with limited labeled data. A diverse dataset of facial images with varying skin tones and conditions was curated. The proposed approach leverages pre-trained deep neural networks and an FSL algorithm based on prototypical networks (PNs) and matching networks (MNs) to address the challenge of limited labeled data. Importantly, this study has significant implications for improving access to dermatological care, especially in underserved populations, as many individuals are unaware of their skin type, which can lead to ineffective or even harmful skincare practices. Our approach can help individuals quickly determine their skin type and develop a personalized skincare routine based on their unique skin characteristics. The results of our experiments demonstrate the effectiveness of the proposed approach. PNs achieved the highest accuracy in the 2-way, 10-shot, 15-query scenario with an accuracy of 95.78 ± 2.79%, while MNs achieved the highest accuracy of 90.33 ± 4.10% in the 2-way, 5-shot, 10-query scenario. In conclusion, this study highlights the potential of FSL and deep neural networks to overcome the limitations of traditional approaches to facial skin analysis, offering a promising avenue for future research in this field.

Facial skin type analysis is a critical task in several fields, including dermatology, cosmetics, and biometrics, and has been the subject of significant research in recent years. Traditional facial skin type analysis approaches rely on large, labeled datasets, which can be time-consuming and costly to collect. This study proposes a novel few-shot learning (FSL) approach for facial skin type analysis that can accurately classify skin types with limited labeled data. A diverse dataset of facial images with varying skin tones and conditions was curated. The proposed approach leverages pre-trained deep neural networks and an FSL algorithm based on prototypical networks (PNs) and matching networks (MNs) to address the challenge of limited labeled data. Importantly, this study has significant implications for improving access to dermatological care, especially in underserved populations, as many individuals are unaware of their skin type, which can lead to ineffective or even harmful skincare practices. Our approach can help individuals quickly determine their skin type and develop a personalized skincare routine based on their unique skin characteristics. The results of our experiments demonstrate the effectiveness of the proposed approach. PNs achieved the highest accuracy in the 2-way, 10-shot, 15-query scenario with an accuracy of 95.78 ± 2.79%, while MNs achieved the highest accuracy of 90.33 ± 4.10% in the 2-way, 5-shot, 10-query scenario. In conclusion, this study highlights the potential of FSL and deep neural networks to overcome the limitations of traditional approaches to facial skin analysis, offering a promising avenue for future research in this field.

Plant and crop production plays a vital role in the overall development of any country. Sugarcane is one of the most produced crops in Australia. Sugarcane plants are vulnerable to various leaf diseases. Early leaf disease detection is crucial for agricultural productivity. Traditional machine learning often relies on handcrafted features, struggles with high-dimensional data, and has scalability issues. Deep learning models overcomes these problems, still there are some challenges, such as the overfitting issue, the need for a large dataset, and less adaptability to new classes. This research addresses the challenge of limited dataset availability of sugarcane leaf diseases by employing few shot learning (FSL) and episodic training. This research proposes a Prototypical Network (PN) based few-shot learning technique to classify sugarcane leaf diseases. The proposed technique uses pretrained deep learning models for feature extraction, such as MobileNet, ResNet, EfficientNet, and PN based FSL to learn from fewer training examples. The experimental results show that the PN with a backbone of ResNet achieved the highest test accuracy of 87 %. The analysis of the experimental results indicates that PN based few shot learning with deep learning models as backbone outperforms standalone pre-trained deep learning models while using limited number of sugarcane images.

In a manufacturing process, data analysis is conducted to identify defective products in real time to lower their massive production and improve the rate of efficient production. In the production process, it is difficult to find defective products mixed with normal products. Therefore, it is necessary to detect defective products generated in the production process and reduce the risk of their production. Consequently, this study proposes the Mask R-CNN-based occlusion anomaly detection method in consideration of the orientation of manufacturing process data. The proposed method uses Mask R-CNN to find abnormal objects, such as occluded objects, in a manufacturing process line. In the manufacturing process, some products are hidden. Accordingly, preprocessing in consideration of multiple orientations is applied to generate data. The generated data is performed to detect occlusions and anomalies using Mask R-CNN. The mean of IoU was compared to evaluate the detection accuracy of YOLO and Mask R-CNN. YOLO showed excellent performance when there was a constant distance and orientation and no occluded object. However, Mask R-CNN performed excellently when there was any occluded object and the orientation was considered. Therefore, for occlusion anomaly detection in a manufacturing process, Mask R-CNN can reduce the production rate of defective products.

SAPENet: Self-Attention based Prototype Enhancement Network for Few-shot Learning

IntroductionKnee injuries, especially Anterior Cruciate Ligament (ACL) tears and meniscus tears, are becoming increasingly common and can severely restrict mobility and quality of life. Early diagnosis is essential for effective treatment and for preventing long-term complications such as knee osteoarthritis. While deep learning approaches have shown promise in identifying knee injuries from MRI scans, they often require large amounts of labeled data, which can be both scarce and privacy-sensitive.MethodsThis paper analyses a hybrid methodology that integrates few-shot learning with federated learning for the diagnosis of knee injuries using MRI scans. The proposed model used a 3DResNet50 architecture as the backbone to enhance both feature extraction and embedding representation. A combined Centralized and Federated Few-Shot Learning Framework is analysed to leverage episodic-intermittent training strategy based on Prototypical Networks. The model is trained incorporating Stochastic Gradient Descent (SGD), Cross-Entropy Loss, and a MultiStep Learning Rate scheduler to enhance few-shot classification. This model also addressed the challenge of limited annotated data ensuring patient data privacy through distributed learning across multiple regions.ResultsThe models performance was evaluated on the MRNet dataset for multi-label classification. In the centralized setting, the model achieved accuracies of 85.3% on axial views, 82.1% on sagittal views, and 71% on coronal views. The propose work attained accuracies as 83% (axial), 83.9% (sagittal), and 65% (coronal), demonstrating the framework’s effectiveness across different learning configurations.DiscussionThe proposed method outperforms in diagnostic accuracy, generalization across MRI planes, and patient privacy via federated learning. However, it faces limitations, including lower coronal view performance and high computational demands due to its complex architecture.

ABSTRACTBreast cancer (BC) detection based on mammogram images is still an open issue, particularly when there is little annotated data. Combining few‐shot learning (FSL) with transfer learning (TL) has been identified as a potential solution to overcome this problem due to its ability to learn from a few examples while producing robust features for classification. The objective of this study is to use and analyze FSL integrated with TL to enhance the classification accuracy and generalization ability in a limited dataset. The proposed approach integrates the FSL models (prototypical networks, matching networks, and relation networks) with the TL procedures. The models are trained using a small set of samples with annotation and can be assessed using various performance metrics. The models were trained and compared to the TL and the state‐of‐the‐art methods regarding accuracy, precision, recall, F1‐score, and area under the ROC curve (AUC). The models proved to be effective when integrated, and the relation networks model was the most accurate, with an accuracy of 95.6% and an AUC of 0.970. The models provided higher accuracy, recall, and F1‐scores, especially in the case of discerning between normal, benign, and malignant cases, as compared to TL traditional techniques and the various recent state‐of‐the‐art techniques. This integrated approach gives high efficiency, accuracy, and scalability to the whole BC detection process, and it has potential for further medical imaging domains. Future research will explore hyperparameter tuning and incorporating electronic health record systems to enhance diagnostic precision and individualized care.

Abstract— In industrial manufacturing processes, ensuring the quality of products is crucial. This paper proposes a system for detecting defects in products using image template matching techniques implemented in MATLAB. The system's primary function is to compare captured images of products with predefined templates to identify potential defects accurately. The method employed in this system is template matching, a well-established approach in image analysis that allows for efficient defect detection. MATLAB, a widely used software tool for image processing, provides the necessary functionality and a robust set of algorithms to implement the proposed system. The experimental results demonstrate the effectiveness of the approach in detecting various types of defects, such as scratches, cracks, and misalignments. This defect detection system offers a reliable and automated solution for improving the efficiency and productivity of manufacturing industries. By enabling early detection and intervention, it contributes to enhancing product quality control and minimizing defective outputs, ultimately leading to cost savings and customer satisfaction. Keywords— Defect Detection; MATLAB; Template Matching; Image Processing

In many small- and medium-sized enterprises (SMEs), defective products are still manually verified in the manufacturing process. Recently, image classification applying deep learning technology has been successful in classifying images of defective and intact products, although there are few cases of utilizing it in practice. SMEs have limited resources; therefore, it is crucial to make careful decisions when applying new methods. We investigated sample size sensitivity to determine the stable performance of deep learning models when applied to the real world. A simple sequential model was constructed, and the dataset was reconstructed into several sizes. For each case, we observed its statistical indicators, such as accuracy, recall, precision, and F1 score, on the same test dataset. Additionally, the loss, accuracy, and AUROC values for the validation dataset were investigated during training. As a result of the conducted research, we were able to confirm that, with 1000 data points or more, the accuracy exceeded 97%. However, more than 5000 cases were required to achieve stability in the model, which had little possibility of overfitting.

Production defects caused by irresistible factors such as process design problems or differences in steel properties in strip production affect the economic benefits of the enterprise and threaten production safety. Traditional defect detection methods are difficult to achieve real-time and high-precision detection, so developing surface defect detection methods based on deep learning is of great significance for strip production. In order to effectively improve the accuracy of the deep learning model in detecting surface defects on hot-rolled strip, in this work we propose a real-time detection model for surface defects on strip steel based on the YOLOv8n model. Firstly, the newly convolutional layer Con5v is designed to replace the original convolutional layer in the neck, and an attention mechanism is added in front of each Con5v to improve the algorithm’s ability to extract small target information. Secondly, an additional set of upsampled feature extraction units is added to the neck in order to enhance the spatial information of the feature map. Subsequently, a set of feature fusion units is incorporated and the convolutional layers in it are improved to provide better feature maps. Thirdly, the number of decoupling detection heads is increased to receive more high-quality features. The final experimental results show that YOLOv8n-GAM (YOLOv8 Nano Model with Global Attention Mechanism) achieves 81.4mAP and 82.0FPS on the NEU-DET dataset and 71.2mAP and 55.0FPS on the GC10-DET dataset, which are 5.7% and 6.9% higher than those of YOLOv8n, respectively. The model proposed in this paper achieves a comprehensive performance improvement in strip steel.

Surface defect detection in industrial environments is crucial for quality management and has significant research value. General detection networks, such as the YOLO series, have proven effective in various dataset detections. However, due to the complex and varied surface defects of industrial products, many defects occupy a small proportion of the surface and fall into the category of typical small target detection problems. Moreover, the complexity of general detection network architectures relies on high-tech hardware, making it difficult to deploy on devices without GPUs or on edge computing and mobile devices. To meet the practical needs of industrial product defect inspection applications, this paper proposes a lightweight network specifically designed for defect detection in industrial fields. This network is composed of four parts: a backbone network, a multiscale feature aggregation network, a residual enhancement network, and an attention enhancement network. The network includes a backbone network that integrates attention layers for feature extraction, a multiscale feature aggregation network for semantic information, a residual enhancement network for spatial focus, and an attention enhancement network for global–local feature interaction. These components enhance detection performance for diverse defects while maintaining low hardware requirements. Experimental results show that this network outperforms the latest and most popular YOLOv5n and YOLOv8n models in the five indicators P, R, F1, mAP@.5, and GFLOPS when used on four public datasets. It even approaches or surpasses the YOLOv8s and YOLOv5s models with several times the GFLOPS computation. It balances the requirements of lightweight real-time and accuracy in the scenario of industrial product surface defect detection.

Surface defect detection in leather manufacturing faces challenges including multi-scale defects, scarce samples, and texture interference. This study proposes an optimized YOLOv11 framework integrating attention mechanisms, cross-scale feature fusion, and few-shot learning. The backbone network employs spatial-channel attention and feature redundancy reduction to enhance defect discrimination. A cross-scale attention mechanism adaptively fuses multi-resolution features for improved small defect detection. A ProtoNet module addresses sample scarcity while ensuring localization precision. Evaluations on industrial leather datasets and public benchmarks demonstrate the model’s effectiveness, achieving 80.8 mAP on leather defects and 78.1 on steel surfaces with 3.02M parameters and real-time inference 70.3 FPS . The framework outperforms conventional methods in accuracy and robustness, offering a practical solution for automated quality inspection in texture-rich industrial scenarios. Our source code is available at https: github.com zuoqiumama DASC-YOLO.git

Surface defect detection is a core component for ensuring the quality of multi-wedge belt products, yet traditional detection methods based on Convolutional Neural Networks (CNNs) suffer from weak global feature extraction and limited multi-scale feature fusion, severely impacting detection efficiency. Meanwhile, the constrained computational resources in industrial scenarios pose critical challenges to balancing model accuracy and speed. To address these issues, we propose a lightweight defect detection model, LCT-YOLO (Lightweight CNN-Transformer YOLO), which integrates the technical advantages of CNN and Transformer while meeting the lightweight deployment requirements of industrial applications. Specifically, a dynamic channel selection feature extraction network is constructed by introducing the partial channel selection module CSP_PTB to enhance the model’s capability of capturing long-distance semantic information. A lightweight multi-scale feature fusion network is developed by designing a multi-scale path aggregation network and integrating the dynamic upsampling module Dysample in the model’s neck, effectively addressing challenges from feature scale variations. A lightweight shared convolution detection head is introduced to tackle the high computational cost of existing model detection heads, optimizing performance and reducing model complexity through parameter sharing strategies. Finally, model pruning is employed to remove redundant parameters for compression, and knowledge distillation is utilized to compensate for accuracy loss. Experimental results show that the algorithm achieves an mAP50 of 94.3% on the dataset, with 0.8 M parameters and 2.0 GFLOPs computation, outperforming the baseline model by 0.9% in mAP50, reducing parameters by 74.2%, computation by 75.3%, and featuring a model size of only 1.9 M, making it highly suitable for edge device deployment. The algorithm surpasses other advanced methods in both accuracy and speed, not only meeting the dual requirements of detection precision and real-time performance in industrial scenarios but also being successfully applied to practical production, providing robust technical support for defect detection and product quality control.