Computer graphics processing and selective visual display system – Computer graphics processing – Attributes
Reexamination Certificate
1999-11-01
2002-03-12
Jankus, Almis R. (Department: 2672)
Computer graphics processing and selective visual display system
Computer graphics processing
Attributes
Reexamination Certificate
active
06356272
ABSTRACT:
TECHNICAL FIELD
The present invention relates to a texture information assignment method of assigning texture information to a shape model of a real object of interest according to an object image obtained by shooting that real object of interest, an object extraction method of extracting an object portion by removing an undesired portion such as the background from the object image, a three-dimensional model generation method of generating a three-dimensional model of an object of interest, and apparatus of these methods.
BACKGROUND ART
In accordance with the development of computer graphics and the like, there has been intensive efforts to provide a system for practical usage in three-dimensional graphics. However, one appreciable problem in accordance with the spread of such a system of practical usage is the method of obtaining shape data. More specifically, the task of entering the complicated three-dimensional shape of an object having a free-form surface or that resides in the natural world into a computer is extremely tedious and difficult.
Furthermore, in reconstructing an object with a computer and the like, it is difficult to express the texture of the surface of the object in a more realistic manner by just simply reconstructing the shape of the object.
Three-dimensional image information can be handled more easily if the shape information and color/texture information can be reconstructed within the computer based on image information that is obtained by shooting an actual object.
In three-dimensional image communication such as by, for example, the Internet, the opportunity of a general user to create a three-dimensional image who is the transmitter of information will increase. Therefore, the need arises for a simple and compact apparatus that produces a three-dimensional image.
(1) Japanese Patent Laying-Open No. 5-135155 discloses a three-dimensional model generation apparatus that can construct a three-dimensional model from a series of silhouette images of an object of interest placed on a turntable under the condition of normal illumination.
According to this three-dimensional model construction apparatus, an object of interest that is rotated on a turntable is continuously shot by a camera. The silhouette image of the object of interest is extracted from the obtained image by an image processing computer. By measuring the horizontal distance from the contour of the silhouette image to the vertical axis of rotation for the silhouette image, a three-dimensional model is generated according to this horizontal distance and the angle of rotation. More specifically, the contour of the object of interest is extracted from the continuously shot silhouette images to be displayed as a three-dimensional model.
FIG. 1
is a diagram representing the concept of assigning texture information to the three-dimensional model generated as described above according to the image information continuously picked up by a camera.
Japanese Patent Laying-Open No. 5-135155 discloses the case of obtaining image information by continuously rotating an object of interest and shooting the same, i.e., obtaining image information in the resolution level of shape recognition with respect to a three-dimensional model of a human figure. More specifically, an image is picked up for every 1° of rotation to obtain 360 images with respect to the object of interest.
For the sake of simplifying the description, the case of shooting an image for every larger stepped angle will be described hereinafter. However, the essence is identical.
Consider the case of picking up a total of n images by rotating an object of interest for every predetermined angle of rotation, as shown in FIG.
1
. In this case, each image information corresponds to the label number of 1, 2, 3 . . . , n.
The object of interest is represented as a shape model (wire frame model) 300 using a polygon (triangular patch). When texture information is to be assigned to shape model 300, color information (texture information) of the image information of a corresponding label number is assigned for each triangular patch according to the direction of the camera shooting the object of interest.
More specifically, based upon the vector towards the target triangular patch from the axis of rotation of shape model 300, the texture information with respect to the triangular patch is captured from the image that has the direction of the shooting direction vector and this vector most closely matched. Alternatively, from the standpoint of intuition, a plurality of lines such as the circles of longitude of a terrestrial globe can be assumed with respect to the surface of the model. Texture information can be captured from the first image information for the triangular patch in the range of 0° to 1×360
°, from the second image information for the triangular patch in the range of 1×360
° to 2×360
°, and so on. This method of capturing texture information will be referred to as the central projection system hereinafter.
The central projection system is advantageous in that image information can be provided in a one-to-one correspondence with respect to each triangular patch or the constituent element forming the shape model (referred to as “three-dimensional shape constituent element” hereinafter), and that this correspondence can be determined easily.
However, the central projection system is disadvantageous in that the joint of the texture is noticeable when the gloss or the texture of the color information is slightly different due to the illumination and the like since the texture information is assigned from different image information (image information of a different label number) to a three-dimensional shape constituent element that is not present within the same range of rotation angle when viewed from the axis of rotation.
Furthermore, a corresponding three-dimensional shape constituent element may be occluded in the image information obtained from a certain direction of pickup depending upon the shape of the object of interest. There is a case where no texture information corresponding to a certain three-dimensional shape constituent element is included in the corresponding image information.
FIG. 2
is a diagram for describing such a situation. In
FIG. 2
, the relationship is shown of the axis of rotation, the cross section of the object of interest and the object image projected in the camera at a vertical plane including the axis of rotation of the object of interest. When the object of interest takes a shape that has an occluded region that cannot be viewed from the camera as shown in
FIG. 2
, the image information picked up from this angle direction is absent of the texture information corresponding to this occluded region. However, texture information of this occluded region can be captured from another pickup direction that has a certain angle with respect to the previous direction of pickup.
(2) As a conventional method, extraction of an object portion from an image of an object can be effected manually using an auxiliary tool. More specifically, the image of an object obtained by shooting the target object together with the background is divided into a plurality of regions. The operator selects the background area in the image of the object to erase the background area using a mouse or the like. However, this method is disadvantageous in that the burden on the operator for the manual task is too heavy.
Another conventional method of object extraction employs the chroma-key technique. More specifically, the portion of the object is extracted from the image of the object using a backboard of the same color. However, this method is disadvantageous in that a special environment of a backboard of the same color has to be prepared.
A further conventional method of object extraction employs the simple difference method. More specifically, difference processing is effected between an object image and a background image in which only the background of the object of interest is shot to obtain the difference. The area ha
Arakawa Tsutomu
Matsumoto Yukinori
Sugimoto Kazuhide
Terasaki Hajime
Cunningham G. F.
Jankus Almis R.
Sanyo Electric Co,. Ltd.
LandOfFree
Texture information giving method, object extracting method,... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Texture information giving method, object extracting method,..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Texture information giving method, object extracting method,... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2822461