Abstract:Visual surface defect inspection provides an important tool for product quality assessment in a wide range of industrial applications. In recent years, diverse convolutional neural network (CNN) models have been devel-oped for the high-accuracy image classification of natural objects. However, it remains a challenging task to develop deep-learning-based approaches for surface defect inspection of industrial products. The main challenge is that surface defects of industrial products occur in low probability, therefore it is impractical to build a large image dataset containing different types of defects to train/fine-tune deep CNN models to achieve satisfactory performance when applied to industrial inspection tasks. To overcome the above-mentioned problem, we build a light-weight defect classification model based on the pre-trained SqueezeNet architecture and present three effective techniques to achieve high-accuracy defect classification using very few collected defective training samples. First, we experimentally demonstrate that it is reasonable to freeze shallower convolutional layers and only fine-tune deeper layers in the last convolutional stage when there are not enough training samples. Second, we integrate batch-size independent Group Normalization (GN) into SqueezeNet to compute stable statistics based on a limited number of defective samples for training the classification model. Third, we add an auxiliary task to perform the similarity comparison of two defective samples during the training phase, significantly increasing the size of the training data and generating more distinctive defect-specific representations. The experimental results on the NEU-CLS dataset and USB-FS dataset verify the effectiveness of the proposed method. Our method can obtain accurate defect classification results when very few defective training samples are provided, i.e., about 97.69% and 82.92% average classification accuracy on the NEU-CLS and USB-FS datasets, respectively, under the 5-shot setting (5 images per defect category).

Enhancing Supply Chain Defect Detection through Cross-Domain Generalization: A Deep Learning Approach

Deep-Learning-Based Lithium Battery Defect Detection via Cross-Domain Generalization

Cross-Domain Defect GAN: Defect Generation for Defect Detection in Industrial Area

Cross-domain few-shot defect recognition for metal surfaces

Few-shot defect detection using feature enhancement and image generation for manufacturing quality inspection

Dual-Path GAN: A Method for Enhancing Small-scale Defect Detection on Metal Images

Coarse-to-Fine Few-Shot Defect Recognition with Dynamic Weighting and Joint Metric

An Effective Industrial Defect Classification Method under the Few-Shot Setting Via Two-Stream Training

A Deep Learning-Based Surface Defect Inspection System Using Multiscale and Channel-Compressed Features

Evaluating Generalization, Bias, and Fairness in Deep Learning for Metal Surface Defect Detection: A Comparative Study

Augmented Hybrid Learning for Visual Defect Inspection in Real-World Hydrogen Storage Manufacturing Scenarios

Defect Transfer GAN: Diverse Defect Synthesis for Data Augmentation

TransferD2: Automated Defect Detection Approach in Smart Manufacturing using Transfer Learning Techniques

A review of deep learning-based approaches for defect detection in smart manufacturing

A domain generalization network for imbalanced machinery fault diagnosis

DG-GAN: A High Quality Defect Image Generation Method for Defect Detection

Incorporating Supervised Domain Generalization into Data Augmentation

Machine Vision Product Appearance Defect Detection Based on Deep Learning

DG2GAN: improving defect recognition performance with generated defect image sample

Cross-Category Mechanical Fault Diagnosis Based on Deep Few-Shot Learning