ALEXANDRIA, Va., Feb. 11 -- United States Patent no. 12,548,316, issued on Feb. 10, was assigned to Merck Sharp & Dohme LLC (Rahway, N.J.).
"Multi-dimension unified swin transformer for lesion segmentation" was invented by Shaoyan Pan (Decatur, Ga.), Yiqiao Liu (Pittsburgh), Antong Chen (Blue Bell, Pa.) and Gregory Goldmacher (Lincoln, Mass.).
According to the abstract* released by the U.S. Patent & Trademark Office: "A system and method of multi-stage training of a transformer-based machine-learning model. The system performs at least two stages of the following three stages of training: During a first stage, the system pre-trains a transformer encoder via a first machine-learning network using an unlabeled 3D image dataset. During a sec...