Image-Based Surface Defect Detection Using Deep Learning: A Review

Bhatt, Prahar M.; Malhan, Rishi K.; Rajendran, P.; Shah, Brual C.; Thakar, Shantanu; Yoon, Yeo Jung; Gupta, Satyandra K.

doi:10.1115/1.4049535

Cited by 152 publications

(50 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Some studies have a different focus from our research. The article [81] investigates supervised and semi-supervised deep learning algorithms. As for the unsupervised aspect, more attention is paid to the different network architectures used by different types of algorithms.…”

Section: Related Workmentioning

confidence: 99%

A Survey on Unsupervised Industrial Anomaly Detection Algorithms

Cui¹,

Liu²,

Lian³

2022

Preprint

View full text Add to dashboard Cite

In line with the development of Industry 4.0, more and more attention is attracted to the field of surface defect detection. Improving efficiency as well as saving labor costs has steadily become a matter of great concern in industry field, where deep learning-based algorithms performs better than traditional vision inspection methods in recent years. While existing deep learning-based algorithms are biased towards supervised learning, which not only necessitates a huge amount of labeled data and a significant amount of labor, but it is also inefficient and has certain limitations. In contrast, recent research shows that unsupervised learning has great potential in tackling above disadvantages for visual anomaly detection. In this survey, we summarize current challenges and provide a thorough overview of recently proposed unsupervised algorithms for visual anomaly detection covering five categories, whose innovation points and frameworks are described in detail. Meanwhile, information on publicly available datasets containing surface image samples are provided. By comparing different classes of methods, the advantages and disadvantages of anomaly detection algorithms are summarized. It is expected to assist both the research community and industry in developing a broader and cross-domain perspective.

show abstract

Section: Related Workmentioning

confidence: 99%

A Survey on Unsupervised Industrial Anomaly Detection Algorithms

Cui¹,

Liu²,

Lian³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…ese high-resolution scanners are very expensive, and because of the high resolution they use, the scanning process of an entire wafer takes an inexcusably long amount of time [6].…”

Section: Introductionmentioning

confidence: 99%

Classification of Silicon (Si) Wafer Material Defects in Semiconductor Choosers using a Deep Learning ShuffleNet-v2-CNN Model

Doss

Ramakrishnan

Kavitha

et al. 2022

Advances in Materials Science and Engineering

View full text Add to dashboard Cite

e silicon wafer is one of the raw materials used to make semiconductor chipsets. Semiconductor failure or dysfunction could be the result of defects in the layers of this material. As a result, it is essential to work toward the development of a system that is both quick and precise in identifying and classifying wafer defects. Wafer map analysis is necessary for the quality control and analysis of the semiconductor manufacturing process. ere are some failure patterns that can be displayed by wafer maps. ese patterns can provide essential details that can assist engineers in determining the reason of wafer failures. In this research, a deep-learningbased silicon wafer defect identi cation and classi cation model is proposed. e main objective of this research is to identify and classify the silicon wafer defects using the wafer map images. is proposed model identi es and classi es the defects based on the wafer map images from the WM-811K dataset. e proposed model is composed of a pretrained deep transfer learning model called Shu eNet-v2 with convolutional neural network (CNN) architecture. is Shu eNet-v2-CNN performs the defects identi cation and classi cation process following the work ow of data preprocessing, data augmentation, feature extraction, and classi cation. For performance evaluation, the proposed Shu eNet-v2-CNN is evaluated with performance metrics like accuracy, recall, precision, and f1-score. e proposed model has obtained an overall accuracy of 96.93%, 95.40% precision, 96.26% recall, and 95.75% F1-score in classifying the silicon wafer defects based on the wafer map images.

show abstract

“…He et al developed a regression-and classification-based framework for generic industrial defect detection [23]. Bhatt et al summarized the recent progress in surface defect detection using DL techniques and concluded that DL has been widely explored for use in the automation of defect detection [24]. However, the concurrent identification approach has limitations in multiple defect detection owing to the need for labeled defect data for training; some defects may not be detected when they are rare or unlabeled and when data on these defects are insufficient.…”

Section: Introductionmentioning

confidence: 99%

Detection of Surface Defects in Ceramic Tiles With Complex Texture

Zhang

Peng

et al. 2021

IEEE Access

View full text Add to dashboard Cite

To solve the problem of false defect detection owing to the interference of the texture attribute of ceramic tiles, a method for detecting surface defects in complex-textured ceramic tiles is proposed. Based on the visual detection principle of ceramic tile surfaces, an image acquisition system is established to obtain the ceramic tile image. After image segmentation and correction, the surface defects are preliminarily detected using a saliency detection method. Then, the image sub-block containing the defect area is cut out for secondary detection. The false defects are eliminated, and the final detection of the ceramic tile surface defects is completed using the defect determination method. The feasibility and effectiveness of the defect detection method are studied via comparative experiments. Experimental results show that the maximum accuracy rate of the proposed defect detection method is 98.75%, which satisfies the actual detection requirements and confirms the practical significance of the proposed method.INDEX TERMS Complex-textured ceramic tiles, defect determination, image acquisition, image sub-block, saliency detection HUAILIANG ZHANG received his Ph.D.

show abstract

Image-Based Surface Defect Detection Using Deep Learning: A Review

Cited by 152 publications

References 57 publications

A Survey on Unsupervised Industrial Anomaly Detection Algorithms

A Survey on Unsupervised Industrial Anomaly Detection Algorithms

Classification of Silicon (Si) Wafer Material Defects in Semiconductor Choosers using a Deep Learning ShuffleNet-v2-CNN Model

Detection of Surface Defects in Ceramic Tiles With Complex Texture

Contact Info

Product

Resources

About