公开(公告)号：JP2010152884A

公开(公告)日：2010-07-08

申请号：JP2009271250

申请日：2009-11-30

Applicant: Intel Corp ,  インテル・コーポレーション

Inventor： WU YI ,  HAUSSECKER HORST ,  KOZINTSEV IGOR

IPC: G06T1/00

CPC classification number: G06K9/6292 ,  G06K9/00664 ,  G10L15/26

Abstract: PROBLEM TO BE SOLVED: To provide and match target objects in images and video based on contextual information (such as location and direction) as parameters showing a person's location and objects in view of a camera. SOLUTION: An image recognition algorithm includes a keypoints-based comparison and a region-based color comparison. The method of identifying a target image using the algorithm includes: receiving an input at a processing device, the input including data related to the target image; performing a retrieving step including retrieving an image from an image database, and, until the image is either accepted or rejected, designating the image as a candidate image; and performing an image recognition step including using the processing device to perform the image recognition algorithm on the target and candidate images in order to obtain an image recognition algorithm output. COPYRIGHT: (C)2010,JPO&INPIT

Abstract translation: 要解决的问题：基于上下文信息（诸如位置和方向）来提供和匹配图像和视频中的目标对象作为显示人的位置和对象的参数。 解决方案：图像识别算法包括基于关键点的比较和基于区域的颜色比较。 使用该算法识别目标图像的方法包括：在处理装置处接收输入，所述输入包括与所述目标图像有关的数据; 执行检索步骤，包括从图像数据库检索图像，并且直到图像被接受或拒绝为止，将图像指定为候选图像; 以及执行图像识别步骤，包括使用所述处理装置对目标和候选图像执行图像识别算法，以获得图像识别算法输出。 版权所有（C）2010，JPO＆INPIT

公开(公告)号：JP2012108961A

公开(公告)日：2012-06-07

申请号：JP2012048628

申请日：2012-03-05

Applicant: INTEL CORP

Inventor： WU YI ,  HAUSSECKER HORST ,  KOZINTSEV IGOR

IPC: G06T7/00

Abstract: PROBLEM TO BE SOLVED: To provide an image and a target object in a video image and to make the image coincide with the target object on the basis of context information (such as positions and directions) being parameters showing the position of a person and an object existing in a visual field of a camera.SOLUTION: Image recognition algorithm includes color comparison based on a key point and color comparison based on a region. A method for identifying a target image by using the image recognition algorithm is a step for receiving an input by a processor, a step in which the input includes data relating to the target image, and a step for executing retrieval step including retrieval of an image from an image database, and designates the image as a candidate image until the image is received or rejected. An image recognition step including use of the processor is executed to obtain an image recognition algorithm output and to execute the image recognition algorithm on the target image and the candidate image.

公开(公告)号：JP2010244534A

公开(公告)日：2010-10-28

申请号：JP2010073982

申请日：2010-03-29

Applicant: Intel Corp ,  インテル コーポレイション

Inventor： HAUSSECKER HORST W ,  GAT YORAM ,  ETTINGER SCOTT M ,  KOZINTSEV IGOR V ,  WU YI ,  NESTARES OSCAR

IPC: G06T19/00

CPC classification number: G06T17/05 ,  H04N5/23229 ,  H04N5/23238

Abstract: PROBLEM TO BE SOLVED: To solve a problem that individual or thumbnail display of digital photographs gives a viewer a very limited sense of what the person who has took the photographs has experienced when the images have been captured. SOLUTION: A method includes steps of: receiving a first image captured by an image capture device and first meta data related thereto, including geographical information for contents of the first image; receiving a second image and second metadata related thereto, including geographical information for contents of the second image; determining a viewpoint of the first image which represents the location and orientation of the image capture device when the first image has been captured; and creating a view including the first image and the second image. The placement of the first image in the view is based on the first metadata and the viewpoint of the first image, and the placement of the second image relative to the first image is based on the second metadata and the viewpoint of the first image. COPYRIGHT: (C)2011,JPO&INPIT

Abstract translation: 要解决的问题：为了解决数字照片的个人或缩略图显示给观众的非常有限的感觉，即拍摄照片的人在捕获图像时经历了这些问题。 解决方案：一种方法包括以下步骤：接收由图像捕获设备捕获的第一图像和与其相关的第一元数据，包括用于第一图像的内容的地理信息; 接收与其相关的第二图像和第二元数据，包括第二图像内容的地理信息; 当所述第一图像被捕获时，确定表示所述图像捕获装置的位置和取向的所述第一图像的视点; 以及创建包括第一图像和第二图像的视图。 第一图像在视图中的放置基于第一图像的第一元数据和视点，并且第二图像相对于第一图像的放置是基于第二元数据和第一图像的视点。 版权所有（C）2011，JPO＆INPIT

公开(公告)号：WO2012040709A3

公开(公告)日：2012-07-05

申请号：PCT/US2011053271

申请日：2011-09-26

Applicant: INTEL CORP ,  CHOUBASSI MAHA EL ,  KOZINTSEV IGOR V ,  WU YI ,  GAT YORAM ,  HAUSSECKER HORST W

Inventor： CHOUBASSI MAHA EL ,  KOZINTSEV IGOR V ,  WU YI ,  GAT YORAM ,  HAUSSECKER HORST W

IPC: H04N5/262 ,  G06T17/00

CPC classification number: G06T7/0081 ,  G06T7/11 ,  G06T11/00 ,  G06T17/00 ,  G06T2207/10028 ,  G06T2210/56

Abstract: Embodiments of the invention describe processing a first image data and 3D point cloud data to extract a first planar segment from the 3D point cloud data. This first planar segment is associated with an object included in the first image data. A second image data is received, the second image data including the object captured in the first image data. A second planar segment related to the object is generated, where the second planar segment is geometrically consistent with the object as captured in the second image data. This planar segment is generated based, at least in part, on the second image data, the first image data and the first planar segment. Embodiments of the invention may further augment the second image data with content associated with the object. This augmented image may be displayed such that the content is displayed geometrically consistent with the second planar segment.

Abstract translation: 本发明的实施例描述了处理第一图像数据和3D点云数据以从3D点云数据中提取第一平面段。 该第一平面段与包括在第一图像数据中的对象相关联。 接收第二图像数据，第二图像数据包括在第一图像数据中捕获的对象。 产生与对象相关的第二平面段，其中第二平面段与第二图像数据中捕获的对象几何一致。 至少部分地基于第二图像数据，第一图像数据和第一平面片段生成该平面片段。 本发明的实施例可以进一步增加具有与对象相关联的内容的第二图像数据。 可以显示该增强图像，使得内容与第二平面段几何一致地显示。

公开(公告)号：EP2880634A4

公开(公告)日：2016-04-13

申请号：EP13824962

申请日：2013-07-24

Applicant: INTEL CORP

Inventor： WU YI ,  SUN WEI ,  CHU MICHAEL M ,  DRESHAJ ERMAL ,  MUSE PHILIP ,  AINSWORTH LUCAS B ,  SHOEMAKER GARTH ,  KOZINTSEV IGOR V

IPC: G06T15/00 ,  G06T7/00 ,  H04N7/18 ,  H04N13/00 ,  H04N13/02 ,  H04N13/04

CPC classification number: H04N13/0477 ,  G06T7/11 ,  G06T7/194 ,  G06T2207/10028 ,  H04N5/2226 ,  H04N5/23219 ,  H04N7/147 ,  H04N13/004 ,  H04N13/0048 ,  H04N13/0059 ,  H04N13/02 ,  H04N13/04 ,  H04N2013/0092 ,  H04N2213/006

Abstract: Generally, this disclosure provides methods and systems for real-time video communication with three dimensional perception image rendering through generated parallax effects based on identification, segmentation and tracking of foreground and background layers of an image. The system may include an image segmentation module configured to segment a current local video frame into a local foreground layer and a local background layer and to generate a local foreground mask based on an estimated boundary between the local foreground layer and the local background layer; a face tracking module configured to track a position of a local user's face; a background layer estimation module configured to estimate a remote background layer; and an image rendering module configured to render a 3D perception image based on the estimated remote background layer, the current remote video frame and the remote foreground mask.

Abstract translation: 通常，本公开通过基于图像的前景和背景层的识别，分割和跟踪的生成的视差效应来提供具有三维感知图像呈现的实时视频通信的方法和系统。 该系统可以包括图像分割模块，其被配置为基于局部前景层和本地背景层之间的估计边界将当前局部视频帧分割成局部前景层和本地背景层并且生成局部前景屏蔽; 面部跟踪模块，被配置为跟踪本地用户的脸部的位置; 配置为估计远程背景层的背景层估计模块; 以及图像渲染模块，被配置为基于估计的远程背景层，当前远程视频帧和远程前景掩码来渲染3D感知图像。

公开(公告)号：EP3241184A4

公开(公告)日：2018-08-01

申请号：EP15875914

申请日：2015-12-02

Applicant: INTEL CORP

Inventor： WU JUNKAI ,  JIANG YOUNG ,  WU YI

IPC: G06T7/00 ,  G06K9/46 ,  G06K9/62 ,  G06T7/593

CPC classification number: G06T7/593 ,  G06K9/4642 ,  G06K9/6211 ,  G06T2207/10012 ,  G06T2207/20021

Abstract: A system, article, and method of feature matching for multiple images.

公开(公告)号：EP2727332A4

公开(公告)日：2015-12-23

申请号：EP12807935

申请日：2012-07-02

Applicant: INTEL CORP

Inventor： WU YI ,  TAKACS GABRIEL ,  EL CHOUBASSI MAHA ,  KOZINTSEV IGOR V

IPC: H04N5/262 ,  G06T17/00 ,  G06T19/00

CPC classification number: G06T19/006

Abstract: Embodiments of the invention relate to systems, apparatuses and methods to provide image data, augmented with related data, to be displayed on a mobile computing device. Embodiments of the invention display a live view augmented with information identifying an object amongst other objects. Embodiments of the invention may utilize other related data, such as 3D point cloud data, image data and location data related to the object, to obtain a specific location of an object within the live view. Embodiments of the invention may further display a live view with augmented data three-dimensionally consistent with the position and orientation of the image sensor of the mobile computing device.

公开(公告)号：DE112016003118T5

公开(公告)日：2018-04-12

申请号：DE112016003118

申请日：2016-07-29

Applicant: INTEL CORP

Inventor： SESHADRINATHAN KALPANA ,  NESTARES OSCAR ,  MEHDIZADEH ALI ,  STEIN MAX T ,  WU YI ,  GRANGER JAMES

IPC: H04N13/20 ,  G06T7/00

Abstract: Systeme und Verfahren zum Bestimmen von Punkt-zu-Punkt-Abständen aus 3D-Bilddaten. Bei manchen Ausführungsformen repräsentieren zwei Messpunkte, die zum Beispiel durch einen Benutzer spezifiziert werden, Endpunkte auf einem Objekt von Interesse innerhalb eines Einzelbildes. Unter der Annahme, dass alle Punkte, die zwischen diesen Endpunkten liegen, auch zum Objekt von Interesse gehören, können zusätzliche 3D-Daten, die mit Punkten assoziiert sind, die entlang einer durch die Messpunkte definierten Messlinie liegen, wirksam eingesetzt werden, um eine robuste Abstandsmessung bereitzustellen. Bei manchen Ausführungsformen wird eine Total-Least-Squares-Anpassung durchgeführt, zum Beispiel über RPCA (Robust Principal Component Analysis), um lineare Strukturen innerhalb des Satzes der 3D-Koordinaten auf der Messlinie zu identifizieren. Bei manchen Ausführungsbeispielen wird die MCD(Minimum Covariance Determinant)-Schätzfunktion der Kovarianzmatrix der Daten für eine hoch robuste Schätzung des Mehrvariablenorts und der Mehrvariablenstreuung berechnet.

公开(公告)号：EP3213256A4

公开(公告)日：2018-04-11

申请号：EP15855673

申请日：2015-08-26

Applicant: INTEL CORP

Inventor： SESHADRINATHAN KALPANA ,  WU YI

IPC: G06T5/50 ,  G06T5/40

CPC classification number: G06T5/40 ,  G06T5/007 ,  G06T5/50 ,  H04N5/23229 ,  H04N2013/0081

Abstract: Global matching of pixel data across multiple images. Pixel values of an input image are modified to better match a reference image with a slope thresholded histogram matching function. Visual artifacts are reduced by avoiding large pixel value modifications. For large intensity variations across the input and reference image, the slope of the mapping function is thresholded. Modification to the input image is therefore limited and a corresponding modification to the reference image is made to improve the image matching. More than two images may be matched by iteratively modifying a cumulative mass function of a reference image to accommodate thresholded modification of multiple input images. A device may include logic to match pixel values across a plurality of image frames generated from a plurality of image sensors on the device. Once matched, image frames may be reliably processed further for pixel correspondence, or otherwise.

公开(公告)号：EP2756682A4

公开(公告)日：2015-08-19

申请号：EP11872318

申请日：2011-09-12

Applicant: INTEL CORP

Inventor： RATCLIFF JOSHUA ,  WU YI ,  CHOUBASSI MAHA EL ,  GAT YORAM ,  SUN WEI VICTORIA ,  SESHADRINATHAN KALPANA ,  KOZINTSEV IGOR V

IPC: H04N13/02 ,  H04N13/00

CPC classification number: G06T7/0081 ,  G06K9/00228 ,  G06K9/342 ,  G06T7/11 ,  G06T7/174 ,  G06T7/194 ,  G06T7/74 ,  G06T19/20 ,  G06T2207/10012 ,  G06T2207/10024 ,  G06T2207/20221 ,  G06T2207/20224 ,  G06T2207/30196 ,  G06T2207/30201 ,  G06T2207/30244

Abstract: Systems, devices and methods are described including receiving a source image having a foreground portion and a background portion, where the background portion includes image content of a three-dimensional (3D) environment. A camera pose of the source image may be determined by comparing features of the source image to image features of target images of the 3D environment and using the camera pose to segment the foreground portion from the background portion may generate a segmented source image. The resulting segmented source image and the associated camera pose may be stored in a networked database. The camera pose and segmented source image may be used to provide a simulation of the foreground portion in a virtual 3D environment.