A leading computer scientist has invented a way for artificial neural networks to recognize objects it sees, even if they're partially obscured. "Machines can readily accomplish tasks we think difficult, but they cannot yet accomplish among the simplest of tasks for humans," said Dr. Stuart H. Rubin of the Naval Information Warfare Center in San Diego. "That's where the game changer lies." On Tuesday, the U.S. Patent and Trademark Office granted Rubin U.S. Patent 10,217,023. The document, which can be downloaded below, describes Rubin's application of cameras connected to an artificial intelligence system, which uses arrays of spatial light modulators to repetitiously examine the camera's imagery and determine if it matches anything in its memory.