US 10979693

Analysis of U.S. Patent 10,979,693: A Stereoscopic 3D Camera for Virtual Reality

Washington D.C. - A detailed analysis of U.S. Patent No. 10,979,693, titled "Stereoscopic 3D camera for virtual reality experience," reveals a method for enhancing virtual reality by capturing and processing three-dimensional video. The patent, issued on April 13, 2021, is assigned to Artificial Intelligence Imaging Association Inc.

The invention, credited to inventors Han Xiong Jin and Adam Rowell, was filed on February 28, 2018. It addresses the growing demand for immersive virtual reality content by outlining a system and method for capturing, correcting, and presenting stereoscopic video.

The abstract of the patent describes a stereoscopic device that captures 3D images and videos with a wide field of view to create an immersive virtual reality experience for the user.

A search of the United States Court of Appeals for the Federal Circuit (CAFC) dockets for 2026 did not yield any public records associated with this patent number. This indicates that, as of this date, no litigation involving this patent has been docketed with the CAFC for the specified year. However, this does not preclude the possibility of ongoing or future litigation in other venues.

Overview of Independent Claims

The patent asserts two independent claims, which form the core of its legal protection. In plain language, these claims are as follows:

Independent Claim 1: This claim describes a method for correcting and mapping stereoscopic video data. The process begins by obtaining a video stream from a stereoscopic camera, along with data about the camera's field of view and the settings of its lenses or sensors. A "reference frame" of video is identified, along with a set of frames immediately preceding and following it. By comparing these surrounding frames, the system can characterize any motion that occurred. This motion data is then filtered to isolate the movement corresponding to the reference frame.

The system then calculates three distinct mathematical matrices. The first is based on the filtered motion. The second matrix represents a mapping of the 3D space into a 2D plane, as if the camera were perfectly still; this is calculated using the camera's focal length and principal point at the moment the frame was captured. For each of the surrounding frames, intrinsic data in the 3D space is estimated. The third matrix is the inverse of the first. Finally, these three matrices are applied together in a single operation to the original reference frame, resulting in a modified, stabilized frame.

Independent Claim 5: This claim focuses on the playback of video and outlines a similar method for video stabilization. It involves obtaining a reference frame and its surrounding frames from a stereoscopic camera. The motion between these frames is analyzed and filtered. A key part of this claim is the calculation of three matrices: one that maps the 3D scene to a 2D plane in the absence of motion (using the camera's focal length and principal point), a second which is the inverse of the first, and a third that is related to the filtered motion. These three matrices are then applied to the reference frame in a combined operation to produce a modified and stabilized frame for playback. This method essentially allows for real-time video stabilization during the viewing process.