The comprehensive intelligent development of the manufacturing industry puts forward new requirements for the quality inspection of industrial products. This paper summarizes the current research status of machine learning methods in surface defect detection, a key part in the quality inspection of industrial products. First, according to the use of surface features, the application of traditional machine vision surface defect detection methods in industrial product surface defect detection is summarized from three aspects: texture features, color features, and shape features. Secondly, the research status of industrial product surface defect detection based on deep learning technology in recent years is discussed from three aspects: supervised method, unsupervised method, and weak supervised method. Then, the common key problems and their solutions in industrial surface defect detection are systematically summarized; the key problems include real-time problem, small sample problem, small target problem, unbalanced sample problem. Lastly, the commonly used datasets of industrial surface defects in recent years are more comprehensively summarized, and the latest research methods on the MVTec AD dataset are compared, so as to provide some reference for the further research and development of industrial surface defect detection technology.

Fabric defect detection is a crucial problem in the production of textile industry. Nowadays, a large number of fabric defect detection algorithms based on deep learning have been proved to have excellent detection results, but the performance of detection often depends on large-scale labeled datasets. However, it is very laborious and costly to label a large number of pixel-level defect samples in industrial production. Although semisupervised detection method can reduce the labeling workload, fabric defect detection is still a challenging task due to the diversity of defects and the low contrast between defects and background. To address these issues, this article proposes a pixel-wise semisupervised fabric defect detection method combined with multitask mean teacher (MT). We focus on establishing a multitask detection model for defect area detection, defect contour detection, and defect distance map detection by leveraging their auxiliary information in better defect segmentation and assign this model as a student network and teacher network. Finally, the multitask supervised loss for labeled images and the multitask consistency loss for unlabeled images to train the network. Experiments on three public fabric defect detection datasets show that our method outperforms the current mainstream pixel-wise segmentation methods with MT, which verifies that the proposed network can effectively detect fabric defects by a small number of labeled samples.

Due to the complex diversity of both fabric texture and defect, fabric defect detection is a challenge topic. However, most of existing defect detection methods can still detect only one type of fabric defects. In order to solve this problem, we propose a universal and adaptive defect detection algorithm based on dictionary learning for detecting various defects of different fabric texture. Firstly, in order to make the image more balanced and improve the detection accuracy, according to the complexity of fabric texture and the degree of brightness and darkness of the background, we segment defect-free image according to a certain size to obtain the image joint matrix. In order to make the algorithm more universal, next we form a random dictionary instead of over-complete dictionary and fixed dictionary by randomly select feature columns from the image joint matrix, which can extract the most important texture and background information of fabric. Then, we use the Orthogonal Matching Pursuit algorithm to obtain the sparse representation of defect-free image. On this basis, we use the K-Singular Value Decomposition algorithm to update the dictionary iteratively to get the final dictionary. Afterward, the defective image is reconstructed with the learned dictionary to obtain the reconstruction error. Finally, the correction coefficient is obtained through reconstruction error analysis, then the corresponding adaptive threshold is used to detect and mark the location of the defects. The performance of the algorithm is evaluated on the fabric database of raw fabric, yarn-dyed fabric and patterned fabric. The experimental results show that the proposed algorithm has good universality, adaptability and a high detection success rate for different types of fabrics. Particularly, the average detection success rate of dark-red fabric and dot-patterned fabric can reach 100%, and the Precision and the Recall of defect detection evaluation index can also reach 100%.

Gesture recognition has been applied in many fields as it is a natural human–computer communication method. However, recognition of dynamic gesture is still a challenging topic because of complex disturbance information and motion information. In this paper, we propose an effective dynamic gesture recognition method by fusing the prediction results of a two-dimensional (2D) motion representation convolution neural network (CNN) model and three-dimensional (3D) dense convolutional network (DenseNet) model. Firstly, to obtain a compact and discriminative gesture motion representation, the motion history image (MHI) and pseudo-coloring technique were employed to integrate the spatiotemporal motion sequences into a frame image, before being fed into a 2D CNN model for gesture classification. Next, the proposed 3D DenseNet model was used to extract spatiotemporal features directly from Red, Green, Blue (RGB) gesture videos. Finally, the prediction results of the proposed 2D and 3D deep models were blended together to boost recognition performance. The experimental results on two public datasets demonstrate the effectiveness of our proposed method.

In the printing industry, defect detection is of crucial importance for ensuring the quality of printed matter. However, rarely has research been conducted for web offset printing. In this paper, we propose an automatic defect detection method for web offset printing, which consists of determining first row of captured images, image registration and defect detection. Determining the first row of captured images is a particular problem of web offset printing, which has not been studied before. To solve this problem, a fast computational algorithm based on image projection is given, which can convert 2D image searching into 1D feature matching. For image registration, a shape context descriptor is constructed by considering the shape concave-convex feature, which can effectively reduce the dimension of features compared with the traditional image registration method. To tolerate the position difference and brightness deviation between the detected image and the reference image, a modified image subtraction is proposed for defect detection. The experimental results demonstrate the effectiveness of the proposed method.

The classification of marine vessels is one of the important problems of maritime traffic. To fully exploit the complementarity between different features and to more effectively identify marine vessels, a novel feature structure fusion method based on spectral regression discriminant analysis (SF-SRDA) was proposed. Firstly, we selected the different convolutional neural network features that better describe the characteristics of ships, and constructed the features based on graphs by the similarity metric. Then we weighed the concatenate multi-feature and fused their structures according to the linear relationship assumption. Finally, we constructed the optimization formula to solve the fusion features and structure by using spectral regression discriminant analyses. Experiments on the VAIS dataset show that the proposed SF-SRDA method can reduce the feature dimension from the original 102,400 dimensions to 5 dimensions, that the classification accuracy of visible images can reach 87.60%, and that that of the infrared image can reach 74.68% at daytime. The experimental results demonstrate that the proposed method can not only extract the optimal features from the original redundant feature space, but also greatly reduce the dimensions of the feature. Furthermore, the classification performance of SF-SRDA also gets a promising result.