Limited number of cases may yield generalizable models, a proof of concept in deep learning for colon histology

• Egile Nagusiak: Lorne Holland (Egilea), Dongguang Wei (Egilea), Kristin A Olson (Egilea), Anupam Mitra (Egilea), John Paul Graff (Egilea), Andrew D Jones (Egilea), Blythe Durbin-Johnson (Egilea), Ananya Datta Mitra (Egilea), Hooman H Rashidi (Egilea)
• Formatua: Liburua
• Argitaratua: Elsevier, 2020-01-01T00:00:00Z.
• Gaiak: article
• Sarrera elektronikoa: Connect to this object online.

Background: Little is known about the effect of a minimum number of slides required in generating image datasets used to build generalizable machine-learning (ML) models. In addition, the assumption within deep learning is that the increased number of training images will always enhance accuracy and...