Wang, J., Chang, R., Lim, M. K., Chong, S. C., Yang, X., & Pahwa, R. S. (2024). End-to-End Fast Segmentation Framework for 3D Visual Inspection of HBMs. In (Editor), 2024 IEEE 26th Electronics Packaging Technology Conference (EPTC). https://doi.org/10.1109/eptc62800.2024.10909770
Abstract:
In this paper, we present an innovative end-to-end system that automates the segmentation of 3D X-ray scans containing HBMs for visual inspection. Our system streamlines the process by taking a 3D X-ray scan as input and delivering a fully segmented output, eliminating the need for manual steps prevalent in current methods. Traditional approaches are limited by the analysis of individual bumps and do not provide an end-to-end inspection of 3D scans [1-5]. Our automated framework significantly reduces the need for manual interventions, taking only about 200 seconds per scan, compared to the traditional pipeline, which takes around 20 minutes. This framework is also highly modular and customizable as it can include any detection or segmentation model. Different AI models could then be used for a specific type of data or bumps and be loaded on the fly. This improvement enhances both efficiency and convenience for industry adoption.
License type:
Publisher Copyright
Funding Info:
This research / project is supported by the A*STAR - Manufacturing, Trade, and Connectivity Programmatic Fund
Grant Reference no. : M23L7b0021
This research / project is supported by the A*STAR - Career Development Fund
Grant Reference no. : C210812046