ALEXANDRIA, Va., Sept. 30 -- United States Patent no. 12,749,310, issued on Sept. 29, was assigned to Amazon Technologies Inc. (Seattle).

"Merging tokens for long-form video understanding" was invented by Jue Wang (Sammamish, Wash.), David Jiatian Fan (New York), Xinyu Li (Sammamish, Wash.), Zhikang Zhang (Bellevue, Wash.) and Seon-Ho Lee (Seattle).

According to the abstract* released by the U.S. Patent & Trademark Office: "Tokens or patches are divided from individual images of long-form video files and merged by transformer models, for the purposes of classifying such video files. Tokens or patches may have any dimensions, and may be partitioned from images of a video file into target tokens and source tokens in any manner, e.g., unifor...