ALEXANDRIA, Va., April 7 -- United States Patent no. 12,597,143, issued on April 7, was assigned to Altos Labs Inc. (Redwood City, Calif.).
"Semantic segmentation of sparse multi-dimensional tensors" was invented by Wilmer Arbelo-Gonzalez (Berkeley, Calif.), Adam Scott Frost (San Francisco) and Valentina Pedoia (San Francisco).
According to the abstract* released by the U.S. Patent & Trademark Office: "One embodiment of the present invention sets forth a technique for training a machine learning model to perform semantic segmentation. The technique includes iteratively extracting, from a set of 3D images, a first set of three-dimensional (3D) patches having a first patch size and a second set of 3D patches having a second patch size, wher...