Year: 2007-08

Company: Gilead Sciences

Liaison(s): Gerard Jensen, Tark Bunch

Testing of pharmaceutical product components in the manufacturing process is required to ensure quality. Automated and trainable machines could eliminate human variability in characterizing components such as glass vials for parenteral dosages, stoppers for use with vials, and other dimensioncritical items. The technology for automated dimensional analysis and defect detection has undergone rapid improvement and development in recent years. Implementation of automated inspection systems could improve efficiency and reduce operating costs at the manufacturing plant of Gilead Sciences in San Dimas. Currently, there are two levels of inspection performed in the industry—100% inspection of all incoming and final products by the manufacturing department, and a separate verification of batch quality by discrete sampling inspection by the quality assurance department. This project was a joint effort between Harvey Mudd College and Keck Graduate Institute. There were two goals of this project. The first objective was to evaluate and make recommendations for the purchase of a manufacturing-scale automated inspection system for parenteral products. The scope of this evaluation included a risk assessment and an economic justification for the recommended system. The second objective was to design a system to detect and quantify cosmetic defects in pills for quality assurance purposes. The developed prototype is able to detect and classify cosmetic defects in pills using image analysis software and serves as a proof-of-concept for the development an automated inspection system for pill defect detection.