US Patent Issued to Singulos Research on June 10 for "Method and apparatus for efficient non-integer scaling in neural network accelerators" (Canadian Inventors)

ALEXANDRIA, Va., June 18 -- United States Patent no. 12,327,326, issued on June 10, was assigned to Singulos Research Inc. (Burnaby, Canada).

"Method and apparatus for efficient non-integer scaling in neural network accelerators" was invented by Michael Scott Lee (North Vancouver, Canada), Scott Chin (Vancouver, Canada), Bradley Quinton (Vancouver, Canada) and Trent McClements (Burnaby, Canada).

According to the abstract* released by the U.S. Patent & Trademark Office: "Processing image data using deep neural networks is critical to many systems that desire to understand objects and their environment using camera sensors. Image scaling is a fundamental processing task required when managing image data. Although it is possible to scale ima...