Image analysis – Image compression or coding – Shape – icon – or feature-based compression
Reexamination Certificate
2002-11-05
2004-03-30
Johnson, Timothy M. (Department: 2625)
Image analysis
Image compression or coding
Shape, icon, or feature-based compression
C382S239000, C382S240000, C382S299000
Reexamination Certificate
active
06714684
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates generally to an image data coding and/or decoding system which can carry out high-efficient coding of picture signals to transmit and store. More specifically, the invention relates to an image data coding system which can code and transmit picture signals to display an image on a liquid crystal display with a small screen which can be built in a wristwatch and so forth.
In the coding of image data used for a visual telephone (TV phone), a television conference and so forth, the image data efficiently compressed utilizing human's visual characteristic are used. The human's visual characteristic with respect to the distortion of a picture utilized here are as follows (see “Image Information Compression”, issued by Japanese Television Society and complied under the supervision of Hiroshi Harashima, page 12).
(1) Frequency Characteristic in Distortion Perception
Distortion varying with elapsed time and distortion with high spatial frequency are difficult to be visible to the naked eye.
(2) Relationship with Pattern of Image
Distortion is easy to be perceived at the flat portion of the image, and difficult to be visible on the contour portion of the image. However, this is the case of a still picture. In a moving picture, the distortion on the contour portion serves as an edge busyness to conversely offend the eye.
(3) Relationship Between Image and Motion
When a picture is moving at a higher speed than a given speed and the user's eyes can not follow its motion, the perception sensitivity to distortion lowers.
(4) Relationship with Switching of Scene
Immediately after a scene has been switched, the distortion is not to be visible to the naked eye if the resolution considerably lowers.
(5) Relationship with Brightness of Screen
The more the screen is dark, the more the picture distortion of the same level is easy to be visible to the naked eye.
(6) Color Signal and Luminance Signal
Since distortion by color signals is more difficult to be visible to the naked eye than that by luminance signals, for example, it is possible to thin out sampled points of the color signals.
In addition, since visual acuity (spatial resolving power) on the peripheral portions of the visual field is worse than that on the central portion thereof under the influence of the distribution of visual receptor cells on retinas, it is necessary for an user to move his eyes (eye movement) in order to obtain information such as shape, structure and detail contents (see “Image Information Compression” issued by Television Society, published by Ohm, page 41). Therefore, to determine the definition of the picture in view of human's visual characteristic is dominated by the movement of human's eye serving as a subjective factor in addition to the resolution of the picture serving as an objective factor.
On the other hand, when a human looks at an object, if the object is small, it is possible to recognize the whole shape and so forth of the object by staring a specific range around a point. However, if the object is large, it is necessary to closely observe a wide range including a large number of points to recognize the whole shape and so forth of the object. When he watches a television receiver, if its screen is large, a large number of closely observed points are distributed in a given range by frequently moving his eyes, but if the screen is small, the range wherein the closely observed points are distributed does not so extend.
It is disclosed in “Estimation Technique of Image Quality and Tone Quality” (edited Television Society and published by Shokodo, page 118) that since the display screen in a high quality television system which rapidly approaches to implementation in recent years is greater than those of current television systems, the closely observed points distributing ranges in these systems are different. FIG. 5.22 on the same page of this paper shows the measured results of proportion of the closely observed points distributing range to the area of the screen when observing a high quality television system and a current television system on a standard observation condition using a program of the same content. This figure is expressed by approximating to an ellipse with three times as large as the standard deviation assuming that the closely observed points lie on a normal distribution in horizontal and vertical directions when the center of the screen is the origin. It is also shown the experimental results that the proportion of the distributing range of the closely observed points to the area of the screen is about 60% in the current television systems, but it reaches about 80% in the high quality television system. That is, as the size of the screen decreases, the proportion of the distributing range of the closely observed points decreases and the range concentrates on the center of the screen. Therefore, since the spatial resolving power of the visual sensation on the peripheral portion of the screen is inferior, the information compression can be efficiently carried out by lowering the spatial resolution or by weighting the assignment of the distortion in preprocessing.
By the way, as a method for efficiently compressing the measure of information using the difference between the visual characteristic at the central portion of the visual field (central vision) and the visual characteristic at the peripheral portion of the visual field (peripheral vision), there is a method disclosed in, for example, “Visual Pattern Image Sequence Coding” (Aug., 1993, IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL.3, NO.4, pp-291-301). In the technique disclosed in this literature, a function relating to the position of radius r from the central point of the screen is derived, and the resolution on the peripheral portion of the screen is lowered using this function.
In addition, as a method for performing the information compression by changing the distribution of the assigned code amount in a visually important region and an unimportant region, there are two methods as follows.
One of the methods has been proposed as applied to a video telephone (Japanese Patent Application Laid-open No. 1-80185 (1989) “Moving Picture Coding Method”). In this method, on the assumption that the closely observed points are concentrated on the face of the opposite party for the telephone conversation, the face region is detected to assign many code amount on the detected face region.
Another method is also applied to a video telephone similar to the aforementioned proposal (Japanese Patent Application Laid-open No. 5-95541 (1993)). Similar to the aforementioned proposal, by detecting the face region to apply a spatial-temporal filtering to a region other than the face, the code amount produced in this region other than the face is decreased, and the code amount assigned in the face region is increased.
Both of these conventional methods pay attention to human's visual characteristic, and provide a natural picture to a person which visually recognizes a reproduced picture, by changing the coded data amount so that the coding data amount in the region in which the closely observed points are concentrated in the distribution of closely observed points, is different from the coding data amount in the region in which the closely observed points are not so concentrated.
As mentioned above, in both of the conventional image data coding methods, the information compression has been efficiently performed using human's visual characteristic by restraining the code amount produced in a visually unimportant region and by increasing the code amount assigned to a visually important region. However, both of the techniques disclosed in the aforementioned two publications only classify the regions in the screen on the basis of the degree of concentration of the distribution of closely observed points, to vary the code amount assigned to each of the regions, and these techniques do not consider human's visual characteristic that
Chujoh Takeshi
Dachiku Kenshi
Ida Takashi
Kikuchi Yoshihiro
Watanabe Toshiaki
Finnegan Henderson Farabow Garrett & Dunner L.L.P.
Johnson Timothy M.
Kabushiki Kaisha Toshiba
LandOfFree
Subband image compression of blocks within a bounded shape... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Subband image compression of blocks within a bounded shape..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Subband image compression of blocks within a bounded shape... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3258832