Image analysis – Image enhancement or restoration – Variable threshold – gain – or slice level
Reexamination Certificate
1999-09-20
2003-09-23
Mehta, Bhavesh M. (Department: 2625)
Image analysis
Image enhancement or restoration
Variable threshold, gain, or slice level
C382S252000, C358S003220, C358S466000
Reexamination Certificate
active
06625327
ABSTRACT:
This application is based on application Nos. 10-288834, 11-200249 and 11-237492 filed in Japan, the contents of which is hereby incorporated by reference.
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to method and apparatus of image processing, and, more specifically, to method and apparatus of image processing capable of performing gradation reducing process in which number of gradations of image data is reduced, by using threshold values.
2. Description of the Related Art
Handling of images in digital manner is currently dominant in the field of image processing. It is often the case that for displaying or outputting digital image, it becomes necessary to display gradations of the image in smaller number of gradation levels, because of restrictions imposed by characteristics of an output device and the like. From the early stages of development, various methods of digital half toning image processing such as binarization, in which gradations are reproduced solely by white and black dots as a pseudo halftone processing, has been studied.
Various methods including ordered dither method and error diffusion method, which are still utilized at present, as well as descendents of these methods have been developed and improved from 1960's. Further, as the hardware of computation has been developed recently, a method of directly performing optimal search for pixel arrangement, such as the method of cost minimization, has been developed.
These methods of half toning have respective advantages and disadvantages in accordance with the objects of use, and therefore various problems and solutions for respective methods have been studied. For example, the ordered dither method is simple and easy to use, while reproduced image quality is not very good. Though load of computation is heavier in the error diffusion method than the dither method, image quality is better.
In the method of directly performing optimal search such as the method of cost minimization, various optimization methods such as neural network, genetic algorithm and simulated annealing are utilized. Adoption of such a method facilitates incorporation of a visual model or an output device model into the process, enlarging degree of freedom in the processing. On the other hand, as the optimal state is searched through repetitive operations, load becomes formidable.
The problems change along with the development of technology. The problem of formidable load experienced when the method of directly performing optimal search is used may be solved by the development of hardware defining the speed of calculation. From the viewpoint of promoting wide spread use of simple and high quality output devices, however, simpler calculation process is desired.
Further, there are the problem of trade off between resolution and gradation common to all the methods. This problem may possibly be solved by increased output gradation levels or improved resolution characteristic of the output device itself. It is expected, however, that there will be increased occasions where characters are processed as images, and such processing should desirably be done in the simplest manner possible.
Conventionally, methods of improving image processing have been studied, including a method in which an image region of which gradation is of importance and an image region of which resolution is of importance are determined and the method of processing is changed in accordance with the result of determination for respective regions, and a method in which a plurality of processing methods are combined. These methods are hardly said to be simple methods, as a new process of region determination, for example, must be developed and added to execute such methods. Considering balance with the hardware (output device), it is desirable that satisfactory resolution and gradation are both attained through such a method that is comparable to the error diffusion method.
FIG. 66
is a block diagram showing a configuration of a conventional image processing apparatus executing the error diffusion method.
Referring to the figure, the image processing apparatus includes: an input unit
501
receiving as an input a pixel value of one pixel of a multi-value image; a subtractor
503
subtracting diffused error from the input pixel value; an output unit
505
outputting, as a corrected pixel value, an output from subtractor
503
; thresholding unit
507
performing thresholding on the output of output unit
505
to provide binary data; an output unit
509
outputting, as pixel data, the output of thresholding unit
507
; a subtractor
511
subtracting the output of output unit
505
from the output of thresholding unit
507
; and an error memory
513
for diffusing the output result from subtractor
511
to pixels around a pixel which is the object of processing (pixel of interest).
The image formed through error diffusion method has a particular texture. The texture is not very noticeable visually, as it has blue noise characteristic. A method of setting dither pattern to attain the blue noise characteristic in a simple manner has been studied for the dither method as well. In the error diffusion method, however, dot patterns are adaptively generated with respect to the input image, and therefore characteristic of the input image is better reflected than the dither method.
In this point, the error diffusion method is superior in image quality to dither method. The error diffusion method, however, has its particular noise. Namely, there occurs a phenomenon in which variation in texture at a region where gradation changes moderately results in an apparent border line where there is no border (texture shift), or a phenomenon in which white or black dots are tend to appear in a line at a region where degradation is close to black or white.
Various methods for improving have been developed to prevent these phenomena, including modulation of weight coefficient and threshold value for error diffusion. As to resolution, though inherent edge enhancement characteristic has been pointed out, it is not sufficient.
Further, from the nature of its algorithm, the error diffusion method functions to reproduce pixel values of the input image in averaging manner. More specifically, the method functions to reproduce local 0th order component of the image. Accordingly, the error diffusion method has been improved to enhance components of 1st and higher order.
SUMMARY OF THE INVENTION
An object of the present invention is to solve the problems of the above described methods of image processing, and to provide apparatus and method of image processing capable of improving image quality.
The above described objects can be attained by an image processing apparatus in accordance with an aspect of the present invention, converting a first image signal representing density level of each pixel in a prescribed number of gradations to a second image signal having a smaller number of gradations than the prescribed number, including a converter successively receiving as inputs first image signals of pixels, comparing density levels of respective pixels with a prescribed threshold and converting to the second image signals, and a feed back circuit based on the signal levels of the second image signals output from the converter and correcting the prescribed threshold value used in subsequent conversion of pixels.
Preferably, the feed back circuit includes control means for controlling a feed back value in the feed back circuit.
Preferably, the control means includes a feed back coefficient setter for setting a feed back coefficient.
Preferably, the feed back coefficient setter is capable of changing feed back coefficient.
Preferably, the feed back coefficient setter sets the feed back coefficient which changes in accordance with density level of each pixel converted by the converter.
Preferably, the feed back coefficient setter includes a calculating unit calculating the feed back coefficient based on a prescribed relation between the feed back coefficient and each density
Ohshima Seiji
Yamamoto Toshitsugu
Bayat Ali
Mehta Bhavesh M.
Minolta Co. , Ltd.
LandOfFree
Method and apparatus of image processing capable of... 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 of image processing capable of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus of image processing capable of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3006604