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).

Center Loss Guided Prototypical Networks for Unbalance Few-Shot Industrial Fault Diagnosis

Industrial fault diagnosis based on few shot learning

Reweighted Regularized Prototypical Network for Few-Shot Fault Diagnosis

Few-Shot Learning for Fault Diagnosis: Semi-Supervised Prototypical Network with Pseudo-Labels

An intelligent fault diagnosis model based on deep neural network for few-shot fault diagnosis

Few-shot Industrial Defect Image Classification Based on Lightweight Model with Attention Mechanism

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

A Novel Imbalanced Data Classification Method Based on Weakly Supervised Learning for Fault Diagnosis

Reweighting and Information-Guidance Networks for Few-Shot Learning

Few-Shot GAN: Improving the Performance of Intelligent Fault Diagnosis in Severe Data Imbalance

An improved prototypical network with L2 prototype correction for few-shot cross-domain fault diagnosis

Selective Kernel Prototypical Network for Few-shot Motor Fault Diagnosis with Unseen Faults

Enhance Prototypical Network with Hybrid Loss for Few-Shot Text Classification

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

Generic Meta-Transfer Learning Model with Special Neuronal Processing Parameters for Few-Shot Fault Bearing Diagnosis

Domain-Adaptive Prototype-Recalibrated Network with Transductive Learning Paradigm for Intelligent Fault Diagnosis under Various Limited Data Conditions

Few-Shot Fault Diagnosis Based on an Attention-Weighted Relation Network

Few-shot transfer learning for intelligent fault diagnosis of machine

Federated Few-Shot Learning-Based Machinery Fault Diagnosis in the Industrial Internet of Things

Few-Shot Learning for Fault Diagnosis With a Dual Graph Neural Network