Recently, the results of the national finals of the 5th China Postgraduate IC Innovation Competition were published. The project Detection System for Multiple Micro-defects of Integrated Circuit Boards Based on High-definition Multi-scale Semantic Information Fusion won the national third prize. The project was under the instruction ofShengli Deng, Professor at our School, designed by the team ofFan Wang, a 2021 grade doctoral student,Sudi Xia, a 2020 grade doctoral student, andQianwen Qian, a 2020 grade postgraduate. This was the first time that the students of our School have won the national prize of this competition.

Group photo of the team, from the left: Qianwen Qian, Fan Wang, Sudi Xia.

The team designed a detection system for multiple micro-defects of integrated circuit boards based on deep learning. The system integrates many functions, such as data enhancement, multi-scale feature fusion, deep learning model training, off-line and on-line identification of various micro-defects of integrated circuit boards, etc. By inputting dynamic data in real time, the system can identify various micro-defects of integrated circuit boards, such as leaking holes, etching gaps, open circuits, short circuits, bumps, pseudo soldering etc., which improves the real-time detection efficiency of integrated circuit board testing by 25%, and provides a new idea for micro-defects detection of large-scale integrated circuit boards.

System introduction of the project.

Fan Wang, the team leader, said that during the preparation, the team tried a variety of different algorithms and combinations of algorithms. Through a large number of experimental results comparison, the detection system with strong portability, rich functions, high accuracy and good real-time performance was finally completed, and two SCI articles and two computer software copyrights related to the competition works were formed. In the next step, the team would make persistent efforts to improve and reconstruct the works, and form an integrated circuit boards test system that integrates component identification and positioning, polarity reverse connection detection and large-scale defect detection, and strive to achieve better results.

China Postgraduate IC Innovation Competition is instructed by the Department of Degree Management and Postgraduate Education of the Ministry of Education, and jointly organized by the Chinese Society of Academic Degrees and Graduate Education and the Youth Science and Technology Center of China Association for Science and Technology. It is one of the competitions of China Postgraduate Innovation Practice Series Competition. China Postgraduate IC Innovation Competition serves the development strategy of the national IC industry, and aims to effectively improve the innovation ability and practical ability of graduate students, and promote the cultivation of outstanding talents in the field of IC. In this Competition, 503 teams from 96 postgraduate education organizations across the country signed up for the competition, and 154 teams were finalists.