Vertical disparity detection in stereoscopic images from optical flow data

    公开(公告)号:US10964034B1

    公开(公告)日:2021-03-30

    申请号:US16668123

    申请日:2019-10-30

    Inventor: David Cook

    Abstract: Due to the factors such as lens distortion and camera misalignment, stereoscopic image pairs often contain vertical disparities. Introduced herein is a method and apparatus that determine and correct vertical disparities in stereoscopic image pairs using an optical flow map. Instead of discarding vertical motion vectors of the optical flow map, the introduced concept extracts and analyzes the vertical motion vectors from the optical flow map and vertically aligns the images using the vertical disparity determined from the vertical motion vectors. The introduced concept recognizes that although not apparent, vertical motion does exist in stereoscopic images and can be used to correct the vertical disparity in stereoscopic images.

    COMPENSATING FOR DISPARITY VARIATION WHEN VIEWING CAPTURED MULTI VIDEO IMAGE STREAMS

    公开(公告)号:US20200020076A1

    公开(公告)日:2020-01-16

    申请号:US16036350

    申请日:2018-07-16

    Inventor: David Cook

    Abstract: The disclosure is directed to a method to compensate for visual distortion when viewing video image streams from a multiple camera capture of a scene where the method determines the disparity difference utilizing the user view orientation and then compresses and/or stretches the left and/or right eye video image streams to compensate for the visual distortion. In another aspect, the method describes additional adjustments and corrections to the video image streams including rotating, tilting, shifting, and scaling the video image streams, and correcting for gapping and clipping visual image artifacts. In another aspect, a visual compensation system is described to implement the method. Additionally, a visual compensation apparatus is disclosed to perform the method operations.

    VERTICAL DISPARITY DETECTION IN STEREOSCOPIC IMAGES USING A DEEP NEURAL NETWORK

    公开(公告)号:US20210183079A1

    公开(公告)日:2021-06-17

    申请号:US17184341

    申请日:2021-02-24

    Inventor: David Cook

    Abstract: Due to the factors such as lens distortion and camera misalignment, stereoscopic image pairs often contain vertical disparities. Introduced herein is a method and apparatus that determine and correct vertical disparities in stereoscopic image pairs using an optical flow map. Instead of discarding vertical motion vectors of the optical flow map, the introduced concept extracts and analyzes the vertical motion vectors from the optical flow map and vertically aligns the images using the vertical disparity determined from the vertical motion vectors. The introduced concept recognizes that although not apparent, vertical motion does exist in stereoscopic images and can be used to correct the vertical disparity in stereoscopic images.

    COLLECTING AND PROCESSING STEREOSCOPIC DIGITAL IMAGE DATA TO PRODUCE A PARALLAX CORRECTED TILTED HEAD VIEW

    公开(公告)号:US20170237963A1

    公开(公告)日:2017-08-17

    申请号:US15043718

    申请日:2016-02-15

    Inventor: David Cook

    Abstract: An apparatus for capturing digital stereoscopic images of a scene. The apparatus comprises a first pair of separated camera lens oriented such that a first imaginary line between the first pair of lens is substantially parallel with a horizon line a scene, wherein digital image data is capturable through the first pair of camera lens and storable in two separate digital image data bases corresponding to a left-eye horizontal view and a right-eye horizontal view respectively. The apparatus comprises a second pair of separated camera lens oriented such that a second imaginary line between the second pair of lens is substantially non-parallel with the horizon line, wherein digital image data is capturable through the second pair of camera lens and storable in two separate digital image data bases corresponding to a left-eye off-horizontal view and a right-eye off-horizontal view respectively.

    GENERATED STEREOSCOPIC VIDEO USING ZENITH AND NADIR VIEW PERSPECTIVES

    公开(公告)号:US20200021790A1

    公开(公告)日:2020-01-16

    申请号:US16528422

    申请日:2019-07-31

    Inventor: David Cook

    Abstract: The disclosure is directed to a method to generate a stereoscopic video image stream of a zenith or nadir view perspective of a scene. In another aspect, a system is disclosed for generating a zenith or nadir view perspective utilizing a relative user view orientation. In yet another aspect, a video processing computer is disclosed operable to generate a zenith or nadir view perspective utilizing a user view orientation.

    Generated stereoscopic video using zenith and nadir view perspectives

    公开(公告)号:US10412361B1

    公开(公告)日:2019-09-10

    申请号:US16036389

    申请日:2018-07-16

    Inventor: David Cook

    Abstract: The disclosure is directed to a method to generate a generated stereoscopic video image stream of a zenith or nadir view perspective of a scene utilizing a blending of stereoscopic and monoscopic view perspectives. In another aspect, a system is disclosed for generating a zenith or nadir view perspective utilizing a relative user view orientation. In another aspect, an apparatus is disclosed capable to generate a zenith or nadir view perspective utilizing a detected user view orientation, and display a generated stereoscopic video image stream of the view perspective.

    SYSTEM AND METHOD FOR MULTI-CLIENT CONTROL OF A COMMON AVATAR
    8.
    发明申请
    SYSTEM AND METHOD FOR MULTI-CLIENT CONTROL OF A COMMON AVATAR 有权
    一种通用AVATAR的多客户端控制系统与方法

    公开(公告)号:US20160287988A1

    公开(公告)日:2016-10-06

    申请号:US15043678

    申请日:2016-02-15

    Abstract: A system for multi-client control of a common avatar. In one embodiment, the system includes: (1) a cloud game engine for executing game code configured to create a game, generate a video stream corresponding to a particular player and accept a response stream from the particular player to allow the particular player to play the game and (2) a cooperative play engine associated with the cloud game engine for communication therewith and configured to multicast the video stream from the cloud game engine to the particular player and at least one other player, combine separate response streams from the particular player and the at least one other player into a joint response stream based on avatar functions contained therein and provide the joint response stream to the cloud game engine.

    Abstract translation: 用于多客户端控制公共化身的系统。 在一个实施例中,系统包括:(1)云游戏引擎,用于执行被配置为创建游戏的游戏代码,生成与特定玩家相对应的视频流,并接收来自特定玩家的响应流,以允许特定玩家播放 游戏和(2)与云游戏引擎相关联的与该游戏引擎相关联的配置游戏引擎,并配置为将来自云游戏引擎的视频流多播到特定玩家和至少一个其他玩家,组合来自特定玩家的单独响应流 并且至少一个其他玩家基于其中包含的头像功能而连接到响应流中,并向云游戏引擎提供联合响应流。

    Reducing video image defects by adjusting frame buffer processes

    公开(公告)号:US10997884B2

    公开(公告)日:2021-05-04

    申请号:US16174600

    申请日:2018-10-30

    Inventor: David Cook

    Abstract: The present disclosure is directed to a method to correct for visual artifacts in a virtual reality (VR) video image where there is significant motion of the video image as a result of user actions. A user may request that the video image be moved, such as a through motion detected through a VR device, i.e., turning the head, or through a request to an application, i.e., joystick feedback to a gaming application. The video image motion can cause stutter and jitter visual artifacts, when the video frame buffer uses a synchronization constraint, such as vertical synchronization (VSync). When the VSync is disabled, a tearing visual artifact can be present. This disclosure presents a frame buffer handling process that operates with VSync disabled. The process allows the display refresh rates to operate at higher frequencies, while correcting for significant motion of the video image, i.e., tearing, through shifting back certain pixels within the scanout frame buffer.

    Hybrid virtual 3D rendering approach to stereovision

    公开(公告)号:US10935788B2

    公开(公告)日:2021-03-02

    申请号:US14604563

    申请日:2015-01-23

    Abstract: A method for stereoscopically presenting visual content is disclosed. The method comprises identifying and distinguishing between a first type of content and a second type of content of a frame to be stereoscopically displayed. The method also comprises rendering the first type of content in a first left and a first right frame from a single perspective using a first stereoscopic rendering method. Further, the method comprises rendering the second type of content in a second left and a second right frame using a second, different stereoscopic method from two different perspectives. Additionally, the method comprises merging the first and second left frames and the first and second right frames to produce a resultant left frame and a resultant right frame. Finally, the method comprises displaying the resultant left frame and the resultant right frame for stereoscopic perception by a viewer.

Patent Agency Ranking