Image analysis – Histogram processing
Reexamination Certificate
1998-08-18
2001-11-20
Lee, Thomas D. (Department: 2722)
Image analysis
Histogram processing
C382S270000, C358S466000
Reexamination Certificate
active
06320981
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates to an image processing system and an image processing method and in particular to an image processing system and its control method for coding and decoding image data efficiently.
In recent years, a digital copier for reading an original image through an image input unit such as a scanner, digitally processing obtained image data, and outputting the result through an image output unit such as a printer, thereby making a hard copy of the original has become pervasive.
The digital copier has an electronic sorter function of storing image data corresponding to a number of images in the copier and sorting originals and an electronic RDH function of filing, page editing, etc., as indispensable functions.
Thus, the digital copier contains memory for temporarily storing image data and a data storage unit such as a large-capacity storage unit typified by a hard disk drive and provides the above-mentioned functions by storing image data and performing processing as required.
To store a large amount of image data, the storage capacity of a data storage unit needs to be increased, but the data storage unit also increases in scale and cost with an increase in the data capacity.
To avoid this problem, a large number of methods for compressing image data and storing the compressed data in a data storage unit are proposed.
Resultantly, a data storage unit having a small capacity can store a large amount of image data.
A laser beam printer using a laser beam is known as an image output unit.
Generally, a page description language is used for a control method of image output to the laser beam printer.
More particularly, a host computer transfers the page description language contents describing text information and image information of the output contents to the laser beam printer instead of transferring the output contents to the printer as a bit map image (raster image). When receiving the page description language, the printer interprets the language contents, expands page image data as a bit map image (raster image), and transfers and outputs the image onto paper.
The printer must contain a function of interpreting the page description language contents and memory for retaining bit map image for expanding image data.
For example, with a monochrome printer using A3-size paper for output, the memory capacity becomes 32 Mbytes when the image output resolution is 400 dpi and output gradation is 256 levels of gray.
Further, since a color printer needs to output four colors of Y, M, C, and K, the memory capacity also becomes four times that of the monochrome printer; when the image output resolution is 400 dpi and output gradation is 256 levels of gray, the memory capacity becomes 128 Mbytes.
Installation of such large-capacity memory in the printer leads to an increase in the scale and cost of the printer; it is not preferred.
Then, to avoid the problem, a method of compressing image data for decreasing the necessary memory capacity is proposed.
To compress image data for decreasing the data capacity, if the number of gradation steps for reproducing an image is decreased, the finally provided image output quality is degraded. The effect is large particularly if an image is stored in a binarized state.
Thus, to provide high-quality image output, it is desirable to store an image in a multiple-valued state.
A large number of methods of compressing multiple-valued image data are available.
A text area and a photo area often are mixed even in a 1-sheet original among originals output to a digital copier or a printer.
A computer-prepared image, so-called computer graphics (CG) image, and a scanned image of a photo, etc., read through a scanner often are mixed in a printer output image.
The CG and scanned images have entirely different image characteristics.
For example, a CG image area in which a CG image is drawn is constant in pixel value change and contains a flat background area with no pixel image change as a large part.
However, the CG image area also contains a text area containing only binary values of monochrome, a gradation area in which pixel values change violently, and the like.
In contrast, a scanned image area in which a scanned image is drawn often contains noise when the image is read through the scanner; pixel values often change finely even in an area which seems to be flat, such as an image background or blank.
Since the CG image and scanned image areas have different image characteristics, proper compression processing to the image characteristics in each of the CG image and scanned image areas needs to be performed to suppress image quality degradation and effectively compress the image data corresponding to an image having mixed CG image and scanned image areas.
To meet such demand, it is necessary to select optimum image processing for each area of the image data corresponding to an image having mixed areas having different image characteristics in response to the image characteristic and perform compression processing accordingly.
A large number of compression techniques for performing optimum image processing for each area are proposed as adaptive image compression technique or multi-mode compression technique.
Since the CG image area often contains an image requiring a high resolution, such as a text or a line drawing, a compression technique suppressing resolution data degradation is desired for the CG image area.
For example, a reversible compression technique of MMR (Modified Read), LZW (Lempel-Ziv-Welch), JBIG (Joint Bi-level Image Group), etc., or a block compression technique of block truncation coding, etc., wherein resolution data is not degraded although gradation data is degraded is appropriate for the CG image area.
Block run length coding of dividing an image into blocks each consisting of n×m pixels and combining the representative value of the block and the number of the continuous blocks having the value is appropriate for a background area with little pixel value change in the CG image area.
Since the scanned image area often contains an image requiring the number of gradation steps rather than resolution, such as a photo or a natural image, a compression technique suppressing gradation data degradation is desired for the scanned image area.
By the way, if the reversible compression technique suppressing image degradation after decompression is applied to the scanned image area, pixel values change violently and entropy is high in the scanned image area, thus the compression efficiency cannot be enhanced in the reversible compression technique.
Then, an irreversible technique capable of retaining gradation data corresponding to the number of gradation steps after decompression is desired for the scanned image area.
For example, ADCT (Adaptive Discrete Cosine Transform), etc., typified by JPEG (Joint Photographic Experts Group) baseline technique adopted as a color fax standardization technique is appropriate for the scanned image area.
The block truncation coding will be discussed.
The block truncation coding is one type of block coding for dividing an image into blocks each consisting of m×n (m and n are natural numbers except m=n=1) pixels, calculating one or more values representing a block for each block, which will be hereinafter referred to as block representative value or values, and to calculate more than one block representative value, determining index numbers indicating which block representative value is assigned to each pixel in the block.
The block truncation coding is a coding technique using block representative values and the index number assigned to each pixel as code data and the provided code data amount and decoded image quality are determined according to the number of block representative values.
A: When the Number of Block Representative Values is One
When the number of block representative values is one, for example, the mean value, etc., in the block is used as the block representative value and is assigned to all pixels in the block.
Resultantly, in the
Fuji 'Xerox Co., Ltd.
Lee Thomas D.
Oliff & Berrigde PLC
LandOfFree
Image processing system and image 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 system and image processing method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing system and image processing method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2569521