Integrated Circuit Packaging Recognition with Tilt Auto Adjustment using Deep Learning Approach

• DOI: 10.33093/jetap.2023.5.2.9
• Publication Date: 2023-09-15T08:09:06Z
• Authors: Siu Hong Loh; Peh Chiong Teh; Jia Jia Sim; Kim Ho Yeap; Yong Kang Lee
• Keywords: Transfer of learning

A deep-learning-based approach for recognizing integrated circuit (IC) packaging type is presented in this paper. The objective of work to design a deep-learning method that can recognize multiple types per detection, performing counting operations, and calculating the centre location an IC with its tilting angle. transfer learning from model You-Only-Look-Once (YOLO) v5 was chosen because it has been trained coco dataset more reliable feature extraction system than other models. In order extract data images, OpenCV used, which allows deep perform efficient analysis input data. Apart that, principal component (PCA) used estimate angle determine rotation each purpose adjustment. developed average confidence score 85% capable operating variety conditions, as demonstrated by ANOVA analysis.