Television – Stereoscopic – Signal formatting
Reexamination Certificate
1996-03-26
2002-05-07
Diep, Nhon T (Department: 2613)
Television
Stereoscopic
Signal formatting
Reexamination Certificate
active
06384859
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to methods for creating an image for a three-dimensional display (hereinafter referred to as a 3-D display), for calculating depth information, and for image processing using the depth information. The method for creating an image for a 3-D display particularly relates to a method for creating a pseudo-viewfinder image shot by a multi-eye camera (a stereo image) from a viewfinder image shot by a monocular camera. The method for calculating depth information relates to a method for obtaining a distance between an object and a viewer, which is applicable to practice the method for creating an image for a 3-D display. The method for image processing using the depth information relates to applications including the creation of an image for a 3-D display and the suspending of creation of an image for a 3-D display, an enhanced display, or the like.
2. Description of the Prior Art
[1] Creation of an Image for a 3-D Display
In fields related to television techniques for creating a 3-D image (a pseudo stereoscopic vision) through the detection of a movement of a 2-D motion image have been known. One typical example of such a technique is a 3-D display employing a time difference method, the principal of which will be described with reference to
FIGS. 1
to
3
.
In a scene where an object moves from left to right while the background stays still, as shown in
FIG. 1
, by reproducing respective images for right and left eyes (hereinafter respectively referred to as right and left eye images) so as to have a predetermined lapse of time between them, as shown in
FIG. 2
, a parallax &thgr; is caused, as shown in FIG.
3
. “A parallax” or “a binocular disparity” is defined as an angular difference between sight vectors directed at one point from right and left eyes, respectively. In
FIG. 1
, since a viewer perceives the car as being closer than the background due to parallax, a pseudo stereoscopic vision is achieved. When the object, the car in this case, moves in the opposite direction, respective images should be reproduced such that the one for a right eye is reproduced earlier than the one for a left eye by a predetermined time, contrary to the example shown in FIG.
3
.
JP Publication No. Sho 55-36240 discloses a display apparatus for a stereoscopic image using depth information, in which the apparatus receives only an image signal shot from one basic direction (that is, a 2-D motion image) among signals from multiple directions and a signal containing the depth information of an object, so as to reproduce within the apparatus the original viewfinder image shot from multiple directions. The purpose of the apparatus is to reduce a transmission bandwidth. The apparatus incorporates a variable delay circuit for causing a time delay while controlling the extent thereof according to depth information. The time delay results in a parallax. According to an output signal from the circuit, image signals are. reproduced for right and left eyes. In this way, a pseudo stereoscopic image is displayed. This publication discloses, as a preferred embodiment of the disclosed apparatus, (1) a device for displaying a pseudo stereoscopic image for a viewer by respectively supplying right and left eye images to two CRT's, which are situated forming a predetermined angle with a hall′ mirror, and (2) a device for displaying a pseudo stereoscopic image for a viewer even if the viewer moves in a horizontal direction, using a lenticular lens fixed to a display screen.
However, the above apparatus works on the condition that depth information is supplied externally. Therefore, if it only receives a 2-D motion image, the apparatus cannot create a pseudo stereoscopic image.
JP Application Laid-Open No. Hei 7-59119 also discloses an apparatus for creating a pseudo stereoscopic image based on a 2-D motion image. The apparatus comprises a detection circuit for detecting a motion vector from a supplied 2-D motion image, and a delay circuit for delaying either a right or a left image according to the motion vector. The delay causes a parallax. This application discloses, as a preferred embodiment of the disclosed apparatus, a head mounted display (HMD), which is a glasses type display for supplying different images to right and left eyes. Through the HMD, a viewer can see a pseudo stereoscopic image.
In this apparatus, however, since the extent of delay is determined according to the magnitude of a motion vector, any object moving at high speed appears to be closer to the viewer, resulting in an unnatural stereoscopic view, which is discordant to the effective distance between the viewer (or the camera) and the object (that is, a depth).
JP Laid-Open Application No. Sho 60-263594 also discloses an apparatus for displaying a pseudo stereoscopic image using a time difference method, in which the apparatus displays right and left images alternatively for every field, so that they are seen alternatively via shutter glasses for every field, as the shutter glasses alternatively open their right and left eyes. This application further discloses a method for generating a stereoscopic effect by providing a longer time difference between right and left images when an object moves at low speed. However, since this apparatus also does not operate based on depth information, it is not really possible for an accurate pseudo stereoscopic image to be created and thus displayed.
“PIXEL” (No. 128), a magazine, issued on May 1, 1993 describes in pages 97 to 102 a pseudo stereoscopic image system using depth information. In the system, an object is first displayed as a gray-scale image where the gray-scale level corresponds to the depth, and then based on the gray level, the. appropriate parallax is calculated in terms of the number of pixels, so that right and left images are created to be seen via shutter glasses. However, the perspective image is manually created and a technique for automating the creation is not disclosed.
National Publication No. Hei 4-504333 (WO88/04804) discloses a method for achieving a pseudo stereoscopic image using depth information. The method comprises steps of dividing a 2-D motion image into some areas, for giving the divided areas depth information, so as to provide each of the areas with a parallax, and for creating a pseudo stereoscopic image. However, the depth information is manually supplied and a technique for automating the supply is not disclosed.
In a research field called “Computer Vision,” a study has been conducted into a method for estimating a 3-D structure and movement of an object. Concretely speaking, the study, which is aimed at self-control of a robot, relates to acquisition of an accurate distance from a viewpoint to an object by either shooting the object using a stereo camera (a multi-eye camera), or using a monocular camera while moving it. Several aspects oil this technology are described in a report, entitled “1990 Picture Coding Symposium of Japan (PCSJ90),” for example, on page 57.
[2] Creation of Depth Information
Computer Vision would enable detection of the depth of an object. However, in the calculation of depth information, which is based on 2-D motion information, suitable images are not always supplied for the calculation. If the calculation is continued even with unsuitable images supplied, serious errors are likely to be caused. That is, if depth information is obtained from such unsuitable images, and then used for the creation of a stereoscopic image, it may be quite likely that the thus created stereoscopic image will be unnatural, i.e., exhibiting such anomalies as a person in the distance appearing closer than a person who actually is closer.
It is to be noted that the idea of obtaining depth information through understanding of a corresponding relationship between frames has been known. For example, JP Application Laid-Open No. Hei 7-71940 (which corresponds to U.S. Pat. No. 5,475,422) mentions, as a prior art, (1) a technique for relating a point
Arakawa Tsutomu
Katayama Masazumi
Matsumoto Yukinori
Sugimoto Kazuhide
Suzuki Osamu
Diep Nhon T
Hogan & Hartson L.L.P.
Sanyo Electric Co,. Ltd.
LandOfFree
Methods for creating an image for a three-dimensional... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Methods for creating an image for a three-dimensional..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Methods for creating an image for a three-dimensional... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2893230