Development of AI-Based Automated Defect Classification
First written in FY2016. 5 of 10 entries say "Advanced", first in FY2018. Latest entry: FY2026.
The whole theme swung hardest in 2022: 1 of 7 companies on the "Worse" side.
5 other companies' latest entry also says "Advanced".
- FY2026AdvancedAnnual report
「The system also offers a range of applications, including defect-type determination, and contributes to CNT pellicle quality control through use by a broad range of customers, from pellicle manufacturers to device manufacturers.」
PELMIS EPM200 enabled automated classification of particles on the front and back surfaces and defect-type determination, expanding the application of AI and automated classification technologies to pellicle quality control. - FY2025AdvancedAnnual report
「This system features high-precision defect classification and judgment functions using our proprietary deep-learning technology, enabling the early detection and analysis of defects that cause device failures.」
Added defect classification and judgment functions using proprietary deep-learning technology to SICA108, expanding the scope of AI-based defect detection and analysis. - FY2023AdvancedAnnual report
「In CIEL, we completely redesigned the optical system of EZ300 to achieve high-sensitivity, high-throughput inspection. As semiconductor devices become finer and more highly stacked, defects originating at wafer edges are increasing. This product meets customers’ need to detect and classify only those defects among the various types that reduce yield. It combines high-precision defect classification using deep-learning technology with the 3D functionality of its proprietary optical system, enabling the acquisition of high-resolution images containing height and depth information and enhancing the ability to extract defects requiring detection.」
Implemented high-precision deep-learning classification in CIEL for AI-based automated defect classification and expanded functionality to include defect extraction using 3D image information. - FY2022AdvancedAnnual report
「We also offer advanced inspection functions, including high-precision image classification, patterned-substrate inspection and extraction of specific types of defects, using AI inspection technology based on deep learning.」
Implemented deep-learning image classification and defect-extraction functions in OPTELICS AI² for AI-based automated defect classification. - FY2021RestatedAnnual report
「In addition, we continue to deepen our precision mechanics technologies for developing high-precision, high-speed stages and our image-processing technologies for defect detection. In recent years, we have also advanced the development of automated defect classification using AI (artificial intelligence), pursuing research and product development that enables us to provide timely solutions to customer needs.」
Continuing to deepen precision mechanics and image-processing technologies and develop automated defect classification using AI. - FY2020RestatedAnnual report
「In addition, we continue to deepen our precision mechanics technologies for developing high-precision, high-speed stages and our image-processing technologies for defect detection. In recent years, we have also advanced the development of automated defect classification using AI (artificial intelligence), pursuing research and product development that enables us to provide timely solutions to customer needs.」
Continuing to deepen precision mechanics and image-processing technologies and develop automated defect classification using AI. - FY2019RestatedAnnual report
「In addition, we continue to deepen our precision mechanics technologies for developing high-precision, high-speed stages and our image-processing technologies for defect detection. In recent years, we have also advanced the development of automated defect classification using AI (artificial intelligence), pursuing research and product development that enables us to provide timely solutions to customer needs.」
Continuing to deepen precision mechanics and image-processing technologies and develop automated defect classification using AI. - FY2018AdvancedAnnual report
「For defect classification, we use a high-speed processing computer and a trained model based on machine-learning algorithms, including the latest deep learning. Based on high-resolution review images produced using the core technologies described above, the system accurately classifies various defects with irregular shapes as they are detected.」
Implemented machine-learning algorithms, including deep learning, in the GALOIS GaN wafer inspection system for AI-based automated defect classification. - FY2017RestatedAnnual report
「In addition, we continue to deepen our precision mechanics technologies for developing high-precision, high-speed stages and our image-processing technologies for defect detection. In recent years, we have also advanced the development of automated defect classification using AI (artificial intelligence), pursuing research and product development that enables us to provide timely solutions to customer needs.」
Continuing to deepen precision mechanics and image-processing technologies and develop automated defect classification using AI. - FY2016First statedAnnual report
「In addition, we continue to deepen our precision mechanics technologies for developing high-precision, high-speed stages and our image-processing technologies for defect detection. In recent years, we have also advanced the development of automated defect classification using AI (artificial intelligence), pursuing research and product development that enables us to provide timely solutions to customer needs.」
Deepening precision mechanics and image-processing technologies while advancing the development of AI-based automated defect classification.