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).

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

Hybrid-Learning-Based Operational Visual Quality Inspection for Edge-Computing-Enabled IoT System

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

Human-machine knowledge hybrid augmentation method for surface defect detection based few-data learning

Human–machine knowledge hybrid augmentation method for surface defect detection based few-data learning

A Learning-Based Approach for Surface Defect Detection Using Small Image Datasets

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

Small-scale defect detection in industrial environment based on lightweight deep learning network

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

Defect classification on limited labeled samples with multiscale feature fusion and semi-supervised learning

Detecting and Classifying Defective Products in Images Using YOLO

Intelligent 3D Visual Defect Detection for Industrial Applications

Survey on Industrial Defect Detection with Deep Learning

Autoencoder-Based Visual Anomaly Localization for Manufacturing Quality Control

Synthetic data augmentation for high-resolution X-ray welding defect detection and classification based on a small number of real samples

ALADA: A lite automatic data augmentation framework for industrial defect detection

Semi-supervised Knowledge Distillation for Tiny Defect Detection

A new method based on YOLOv5 and multiscale data augmentation for visual inspection in substation

Industrial Product Surface Anomaly Detection with Realistic Synthetic Anomalies Based on Defect Map Prediction

Yolo-sd: simulated feature fusion for few-shot industrial defect detection based on YOLOv8 and stable diffusion

GH-QFL: Enhancing Industrial Defect Detection Through Hard Example Mining