Traditional automatic visual inspection (AVI) can yield false rejections up to 20 percent, by some estimates, and “sometimes even more,” according to Jorge Delgado. Delgado, a Process Development Scientific Director at Amgen, is exploring how false rejection rates can be reduced with the addition of deep learning, a type of AI.
Delgado recently presented on this topic at the 2025 ISPE Aseptic Conference. In the article that follows, Delgato explains deep learning, outlines how the regulatory framework supports deep learning AVI implementation, elaborates on the steps necessary for implementation, and weighs in on the evolving role of human inspectors in deep learning in AVI implementation.
Learn more about ISPE conferences.
Pharmaceutical Online is part of the Life Science Connect media group. The publication provides in-depth articles intended for professionals who are actively involved in the development, manufacture, packaging, and distribution of small molecule products that have been approved by the US Food and Drug Administration (US FDA).