Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
1999-12-06
2002-04-23
Le, Vu (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
C348S043000
Reexamination Certificate
active
06377625
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method for generating a stereoscopic image and an apparatus therefor, and more particularly, to a method for generating a stereoscopic image efficiently and an apparatus therefor, in which the stereoscopic image having a natural and stable stereoscopic sense which is adaptive to a variety of motion types of a picture is generated and a motion vector of each of I, P and B pictures is extracted and generated from a MPEG (Moving Picture Experts Group) bitstream.
Although a stereoscopic camera adopting two cameras is not used, the present invention enables two-dimensional motion images corresponding to MPEG-1 CD (compact disc), MPEG-2 HDTV (high definition TV) or MPEG-2 DVD (digital versatile disc) to be viewed directly as stereoscopic images. Furthermore, commercial applications of the present invention cover all fields that enable video pictures to be viewed as stereoscopic images. In more particular, the present invention can be applied to analog type video signal fields such as NTSC, PAL and SECAM broadcasting systems as well as all fields using MPEG data.
2. Description of the Related Art
As one prior art technology compared with the present invention, a modified time difference (hereinafter referred to as MTD) method has been developed by SANYO DENKI Co., Ltd., in Japan. Also, a method of converting a vertical parallax into a horizontal parallax has been developed by SAMSUNG ELECTRONICS Co., Ltd., in Korea. These prior art technologies generate stereoscopic images from analog type image signals. The MTD method developed by SANYO DENKI Co., Ltd. 1) extracts a moving area, 2) determines a velocity and direction of motion, 3) determining a delay direction and a delay time from a previous frame based on the velocity and direction of motion, and 4) determining whether a delay image delayed according to the determined delay direction and delay time is viewed through which eye of the right and left eyes.
The method of converting a vertical parallax into a horizontal parallax developed by SAMSUNG ELECTRONICS Co., Ltd., 1) predicts a motion block by block, 2) calculates a magnitude of a block motion vector, 3) converts the magnitude value obtained by 2) into a horizontal parallax value in order to convert all vertical components into horizontal parallax values, 4) moves each block by the horizontal parallax values in the horizontal direction, and 5) generates a mixed picture by gathering the horizontally moved blocks.
However, since one of the previous K frames is selected as a delay image forcedly by motion information in the case of the above-described MTD method, different depths possessed by areas in one frame are neglected. As a result, since a sense of depth does not exist when a stereoscopic image is viewed, a stereoscopic sense exists with respect to a moving object but does not exist with respect to a portion having less movement, for example, a background of a picture.
In the case of the method proposed by SAMSUNG ELECTRONICS Co., Ltd., a distortion of an image occurs due to a horizontal movement of each block in accordance with a difference of a horizontal parallax between neighboring blocks. Accordingly, a natural and stable stereoscopic image cannot be generated adaptively to a variety of motion types of an image.
Thus, in the case of the stereoscopic images generated by the above-described existing methods, a stereoscopic sense cannot be obtained in particular motion types of an image, or a motion of the stereoscopic image is not natural.
In addition, the above-described existing methods should perform a motion prediction method again with respect to a decoded image when a stereoscopic image is generated using MPEG encoded data. Accordingly, motion information contained in a MPEG bitstream is not used. As a result, algorithm and hardware for generating steroscopic images becomes complicated due to the reuse of the motion prediction method.
SUMMARY OF THE INVENTION
To solve the above problems, it is an object of the present invention to provide a method for generating an optimal stereoscopic image having a natural and stable stereoscopic sense even in a variety of motion types of an image.
It is another object of the present invention to provide an apparatus for generating an optimal stereoscopic image having a natural and stable stereoscopic sense even in a variety of motion types of an image.
It is still another object of the present invention to provide a method for efficiently generating an optimal stereoscopic image having a natural and stable stereoscopic sense even in a variety of motion types of an image, by using encoded MPEG data.
It is yet another object of the present invention to provide an apparatus for efficiently generating an optimal stereoscopic image having a natural and stable stereoscopic sense even in a variety of motion types of an image, by using encoded MPEG data.
To accomplish the object of the present invention, there is provided a stereoscopic image generation method for generating a stereoscopic image from image data, the method comprising the steps of: (a) determining a motion type of a current picture into one of a still picture, a non-horizontal motion picture, a horizontal motion picture and a fast motion picture, using motion vectors in each picture according to the image data; (b) generating a mixed picture corresponding to the current picture using previous pictures according to the image data and the determined motion type and outputting the current picture and the mixed picture; and (c) receiving the output current picture and the mixed picture and determining the current picture into one of the left picture and the right picture forming a stereoscopic image and the mixed picture into the other thereof, in accordance with the motion type determined in step (a).
To accomplish another object of the present invention, there is also provided a stereoscopic image generation apparatus for generating a stereoscopic image from image data, the apparatus comprising: a motion type determiner for determining a motion type of a current picture into one of a still picture, a non-horizontal motion picture, a horizontal motion picture and a fast motion picture, using motion vectors in each picture according to the image data; a stereoscopic image generator for generating a mixed picture corresponding to the current picture using previous pictures according to the image data and the determined motion type and outputting the current picture and the mixed picture; and a left/right picture determiner for receiving the current picture and mixed picture output from the stereoscopic image generator and determining the current picture into one of the left picture and the right picture forming a stereoscopic image and the mixed picture into the other thereof, in accordance with the motion type determined in the motion type determiner.
To accomplish still another object of the present invention, there is also provided a stereoscopic image generation method for generating a stereoscopic image using MPEG encoded data containing I, B and P pictures, the method comprising the steps of: (a) receiving the encoded picture data and generating and storing motion vectors corresponding to macroblocks in the picture; (b) storing pictures reproducted from the encoded picture data by the number K of pictures; (c) determining a motion type corresponding to a current picture using the motion vectors stored in step (a); (d) generating a mixed picture corresponding to the current picture using the previous pictures stored in step (b) and the motion type determined in step (c) and outputting the current picture and the mixed picture; and (e) receiving the current picture and the mixed picture output from step (d) and determining the current picture into one of the left picture and the right picture forming a stereoscopic image and the mixed picture into the other thereof.
To accomplish yet another object of the present invention, there is also provided a stereoscopic image generation appar
Buczynski Joseph J.
Le Vu
Roylance Abrams Berdo & Goodman L.L.P.
Soft4D Co., Ltd.
LandOfFree
Method and apparatus for generating steroscopic image using... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for generating steroscopic image using..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for generating steroscopic image using... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2933454