Image analysis – Image enhancement or restoration – Object boundary expansion or contraction
Reexamination Certificate
2001-02-12
2004-08-10
Mehta, Bhavesh M. (Department: 2625)
Image analysis
Image enhancement or restoration
Object boundary expansion or contraction
C382S259000, C382S266000, C382S275000, C358S003260, C358S003270, C345S611000, C345S616000, C345S618000
Reexamination Certificate
active
06775418
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to an image processing device and an image processing method, and is particularly characterized by its thinning-out means used in image resolution conversion.
BACKGROUND OF THE INVENTION
Recently, digital-imaging/digital-video equipment, such as a digital still-video camera and DVD, and especially multi-media related equipment among them is rapidly springing into wide use. Also in the field of display devices, dot-matrix displays, such as liquid crystal displays and plasma displays, are becoming widespread. Then, when a number of pixels of image data differs from a number of pixels of a device for displaying it, image processing for adjusting the number of pixels of the image data to those of the display device, i.e. resolution conversion of the image data, is required.
In order to display pictures in a system in which conventional display device having its own number of pixels display picture signal after thinning out some pixels, the method disclosed in Japanese Patent Application Publication No. H05-276436, a technique of detecting correlation between pixels and determining pixels to be thinned out, mean value reducing processing, or the like, has been used. Conventional thinning-out techniques in image processing are described below with reference to the block diagrams shown in
FIGS. 5 and 6
.
In
FIG. 5
, delay circuit
51
delays input signal and feeds the delayed pixels into image discriminating circuit
52
. Image discriminating circuit
52
detects a correlation of difference between adjacent target pixels, and determines pixels with a smaller difference therebetween as pixels to be thinned out. It also determines if the pixels are a part of characters or a part of a natural image. When the pixels have a correlation of a small difference in gray level, they are determined as a part of a natural image and fed into linear interpolation circuit
53
. When the pixels have a correlation of a large difference in gray level, they are determined as a part of characters and fed into adaptive thinning-out circuit
54
. The signal thinned out by both processing circuits are reorganized by image reorganizing circuit
55
as an image and sent to respective pixels in a display device and displayed there.
FIG. 6
is a block diagram of a system using mean value reducing processing. Delay circuit
61
delays input signal and feeds the delayed pixels into operation circuit
62
. Operation circuit
62
averages the pixel to be thinned out and the pixel adjacent thereto, extracts the pixel at a timing of a thinning-out pulse, and feeds the signal obtained after the thinning-out processing into image reorganizing circuit
63
. Image reorganizing circuit
63
reorganizes the picture signal obtained after the thinning-out processing as an image and supplies the data to a display device at a timing corresponding to each pixel in the display device. The signal supplied from image reorganizing circuit
63
are sent to the respective pixels in the display device and displayed there.
FIGS. 7A and 7B
show a case where a black-and-white image based on an amount of unit information (pixels) and including patterns of 8 lines and 8 columns is input. They show 8×8 pixels before the thinning-out processing (FIG.
7
A), and 6×6 pixels obtained after the system shown in
FIG. 6
has performed the thinning-out processing (FIG.
7
B). In the 8×8 pixels, those in columns b, c and f, g, and those in lines b, c and f, g are a pixel to be thinned out and a pixel adjacent thereto, respectively, and pixels B and F are produced from the pixels to be thinned out and the pixels adjacent thereto by the mean value reducing processing. In this drawing, display “H” can be read as “H” after the processing. Although the mean value reducing processing has caused no lack of information, it has produced gradation in the part on which it has been performed. It is because the reducing processing has reduced the image by simply adding up the pixel to be thinned out and the pixel adjacent thereto, i.e. those in columns b, c and f, g, and those in lines b, c and f. g, and diving the sum into equal halves to produce converted pixels.
FIGS. 8A and 8B
show 4×4 pixels before the thinning-out processing (
FIG. 8A
) and 3×3 pixels obtained after the system using the mean value reducing processing shown in
FIG. 6
has performed thinning-out processing (FIG.
8
B). In the case shown in
FIGS. 8A and 8B
, the pixels to be thinned out and the pixels adjacent thereto, i.e. those in columns b and c and those in lines b and c, respectively, are composed of black and white. The mean value reducing processing cannot represent the original black line and produces an indistinct gray line; thus making the reorganized image more blurred than its original. With this processing technique, the reorganized line is more blurred than that in
FIGS. 4A and 4B
obtained by the processing method of the present invention described afterwards. With this processing method, lines and texts shown on a screen generally supplied by a computer are quite blurred.
Next, a case where a conventional thinning-out technique is used for a natural image is described. In
FIGS. 9A and 9B
, thinning-out processing is performed on input picture signal of a natural image by the system using the mean value reducing processing shown in FIG.
6
.
FIGS. 9A and 9B
show input picture signal of a natural image, or the like, with gradation before the thinning-out processing that consist of input signal a, b, c, d, e, f, g, h, i, j, and k arranged in order, and a signal waveform that consists of signal a, B, d, e, f, G, i, j, and k and is obtained after the system using the mean value reducing processing shown in
FIG. 6
has reduced five pixels (
FIG. 9A
) into four pixels (FIG.
9
B). The numerical values are given as a guide of signal levels. In
FIGS. 9A and 9B
, when the mean value reducing processing is performed on pixel b to be thinned out and pixel c adjacent thereto, thinned out pixel B is produced with a signal level of 15 according to (10+20)/2. This thinning-out processing is performed on a portion of the stepped input signal that has a small variation in brightness. The linearity has slightly been lost; however, it is not such a level that causes a problem in the output signal obtained after the processing. Subsequently, when the mean value reducing processing is performed on a pixel g to be thinned out and a pixel h adjacent thereto, thinned out pixel G is produced with a signal level of 50 according to (60+40)/2. This thinning-out processing is performed on the peak of the stepped input signal and the peak of this input signal has been lost. That is, the output waveform after the processing has no edge, thus giving a blurred impression. Therefore, the final image becomes indistinct. This phenomenon is more pronounced as the thinning-out ratio is larger. In the above description, the thinning-out processing in the horizontal direction is described as an example, and in the vertical direction also, similar output signal can be obtained after the processing with completely the same operation, except that the direction in which pixels are thinned out is different.
However, when a natural image is thinned out using the conventional thinning-out processing shown in
FIG. 5
, image discriminating circuit
52
may make a mistake. If it determines a natural image as a part of characters by mistake, straight and curved lines outlining an object are output as irregular lines; thus a more unnatural image than its original is displayed. When the thinning-out processing is performed on an image composed of characters like texts displayed on a personal computer, the same seen in a natural image holds true. In addition, the conventional system has a problem that the circuitry for thinning out characters, and the like, is complicated, and moreover, a large amount of operations are required for thinning-out processing because thinning out method must be changed according to discrimination
Tatsukawa Kouji
Yamauchi Toshiyuki
Yamazaki Kouichi
Kassa Yosef
Matsushita Electric - Industrial Co., Ltd.
Mehta Bhavesh M.
RatnerPrestia
LandOfFree
Image processing device and processing 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 Image processing device and processing method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing device and processing method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3288405