公开(公告)号：EP1498847A3

公开(公告)日：2006-02-15

申请号：EP04076850.9

申请日：2004-06-25

Applicant: STMicroelectronics S.r.l.

Inventor： Curti, Salvatore ,  Battiato, Sebastiano ,  Scordato, Emiliano ,  Tortora, Marcello ,  Ardizzone, Edoardo ,  La Cascia, Marco

IPC: G06T7/00

CPC classification number: G06T7/536

Abstract: There is described a method of obtaining a depth map from a single digital image of a three-dimensional scene containing geometric objects. It envisages the identification in the image of at least one horizontal plane or one vertical plane of the three-dimensional scene and the assignment to every point of the image forming part of the identified plane, or to each one of the identified planes, of a depth level according to its distance from, respectively, a predetermined horizontal or vertical reference straight of the image. In particular, the identification of at least one plane comprises the identification in the image of at least two vanishing lines corresponding to two parallel lines in the plane and the assignment of a depth level is made in accordance with the angular coefficients of the identified vanishing lines.
Also described is a completely automatic method for obtaining a depth map from a single digital image of any three-dimensional scene.

公开(公告)号：EP1507232A2

公开(公告)日：2005-02-16

申请号：EP04015830.5

申请日：2004-07-06

Applicant: STMicroelectronics S.r.l.

Inventor： Curti, Salvatore ,  Battiato, Sebastiono ,  Scordato, Emiliano ,  Tortora, Marcello ,  Ardizzone, Edoardo ,  La Cascia, Marco

IPC: G06T5/00

CPC classification number: G06K9/00704 ,  G06F17/3025 ,  G06K9/00697 ,  G06T7/11

Abstract: A method (Img_Class) for associating with a digital image (Img RGB ) a class of a plurality of predefined classes characterized by respective models. The method comprises the following phases:

dividing (Reg_Detect) the digital image pixel by pixel into one or more regions belonging to a set of predefined regions that differ from each other on account of their type of content, the division being effected by establishing whether or not a pixel of the image belongs to a respective region on the basis of an operation of analyzing the parameters this pixel, the analysis operation being carried out by verifying that the parameters satisfy predefined conditions and/or logicomathematical relationships of belonging to the respective region,
acquiring (f 1 ) from the digital image divided into regions (Img Reg ) information regarding the regions that are present in it,
comparing (Cmp) this information with at least one model characterizing a respective class of said plurality,
associating (SetCL 1 ) with the digital image a class (I_Class) on the basis of the comparison phase.

公开(公告)号：EP1353518A1

公开(公告)日：2003-10-15

申请号：EP02425218.1

申请日：2002-04-09

Applicant: STMicroelectronics S.r.l.

Inventor： Curti, Salvatore ,  Sirtori, Daniele

IPC: H04N13/00

CPC classification number: H04N13/261 ,  H04N13/122 ,  H04N13/128 ,  H04N13/239 ,  H04N13/286 ,  H04N2013/0081

Abstract: To generate, from at least on monocular image (12), at least one image with stereoscopic effect (18, 20), a depth map is provided (14) of said monocular image (a). The depth map in question defines a mapping of the objects present in the monocular image in relation to their spatial position in the direction of depth. For the pixels included in the monocular image and as a function of the depth map, respective parallax values are determined (16) indicative of a respective quantity of movement by which each pixel must be moved in the image with stereoscopic effect (18, 20) in relation to the monocular image (12). A movement corresponding to the respective quantity thus defined is then applied to the pixels of the monocular image (12), if necessary dividing it between two images of a stereoscopic pair (18, 20). Post-processing (22) of the image generated is also envisaged, to eliminate artifacts such as superimposed areas between moved pixels, a lateral edge of the image or gaps between objects included in said image generated as a result of the aforesaid movement of pixels.

Abstract translation: 为了从至少在单眼图像（12）上产生具有立体效果的至少一个图像（18,20），提供所述单目图像（a）的深度图（14）。 所讨论的深度图定义了相对于其在深度方向上的空间位置的存在于单目图像中的物体的映射。 对于包括在单目图像中的像素和作为深度图的函数，确定各自的视差值（16），其指示每个像素必须以立体效果在图像中移动的相应的移动量（18,20） 相对于单目图像（12）。 如果需要，将对应于如此定义的相应数量的运动应用于单目图像（12）的像素，如果需要，则将它分割成立体对（18,20）的两个图像。 还设想了生成的图像的后处理（22），以消除诸如移动的像素之间的叠加区域，图像的横向边缘或由于像素的上述移动而产生的所述图像中包括的对象之间的间隙之间的假象。

公开(公告)号：EP1498847A2

公开(公告)日：2005-01-19

申请号：EP04076850.9

申请日：2004-06-25

Applicant: STMicroelectronics S.r.l.

Inventor： Curti, Salvatore ,  Battiato, Sebastiano ,  Scordato, Emiliano ,  Tortora, Marcello ,  Ardizzone, Edoardo ,  La Cascia, Marco

IPC: G06T7/00

CPC classification number: G06T7/536

Abstract: There is described a method of obtaining a depth map from a single digital image of a three-dimensional scene containing geometric objects. It envisages the identification in the image of at least one horizontal plane or one vertical plane of the three-dimensional scene and the assignment to every point of the image forming part of the identified plane, or to each one of the identified planes, of a depth level according to its distance from, respectively, a predetermined horizontal or vertical reference straight of the image. In particular, the identification of at least one plane comprises the identification in the image of at least two vanishing lines corresponding to two parallel lines in the plane and the assignment of a depth level is made in accordance with the angular coefficients of the identified vanishing lines.
Also described is a completely automatic method for obtaining a depth map from a single digital image of any three-dimensional scene.

Abstract translation: 描述了从包含几何对象的三维场景的单个数字图像中获得深度图的方法。 它设想在三维场景的至少一个水平面或一个垂直平面的图像中进行识别，以及对形成所识别平面的一部分或者被识别平面中的每一个的图像的每个点的分配 根据其与图像的预定水平或垂直参考直线的距离的深度级别。 特别地，至少一个平面的识别包括在平面中对应于两个平行线的至少两个消失线的图像中的识别，并且根据所识别的消失线的角度系数进行深度级的分配 。 还描述了从任何三维场景的单个数字图像获得深度图的完全自动的方法。