Image analysis – Image segmentation – Distinguishing text from other regions
Reexamination Certificate
1999-12-10
2004-04-06
Mehta, Bhavesh M. (Department: 2625)
Image analysis
Image segmentation
Distinguishing text from other regions
C382S177000, C382S261000, C358S462000, C358S453000
Reexamination Certificate
active
06718059
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to systems for processing pixel data. More specifically, the present invention concerns image processing systems in which input pixel data is processed based on detected characteristics of the pixel data.
2. Description of the Related Art
Conventional image processing systems such as a color copier obtain pixel data by scanning an original document, and perform various image processing steps on the data to produce output data suitable for delivery to a laser beam reproduction engine, an ink jet reproduction system, or the like. In particular, conventional image processing systems, such as that in U.S. patent application Ser. No. 08/954,226, which is a continuation application of U.S. patent application Ser. No. 08/365,678, entitled “Image Processing Apparatus And Method”, first receive red (R), green (G) and blue (B) signals representing a scanned image from a charge-coupled device (CCD). The received signals are then subjected to image processing.
FIG. 1
is a diagram illustrating such image processing. As shown in
FIG. 1
, image signals output from a CCD are input to analog signal processing unit
101
, wherein the signal is processed with gain and offset adjustment. Next, each of the R, G and B signals is converted into an 8-bit digital image signal, R
1
, G
1
, and B
1
, respectively, by A/D converter
102
. These signals are then input to shading correction circuit
103
for application of shading correction to each signal. Line delay circuits
104
and
105
are used to compensate for spacing of sensors within the CCD so as to match timing between each of the R
1
, G
1
and B
1
signals such that, after line delay circuit
105
, values of the R, G and B signals at a same point in time represent a same pixel.
Input masking unit
106
converts a reading color space, determined by color decomposition characteristics of the CCD, into a standard color space, and log converter
107
converts luminance signals R
4
, G
4
and B
4
into density signals C
0
, M
0
and Y
0
. The density signals are delayed by line delay memory
108
until determination signals UCR (under color removal), FILTER and SEN can be generated.
After delay of the signals by line delay memory
108
, masking UCR circuit
109
extracts black signals from the density signals using the UCR signal and variable magnification circuit
110
expands and compresses an image signal and a black character determination signal in the main scanning direction. Space filter processing unit
111
performs filtering using the FILTER signal and the resulting frame-sequential image signals M
4
, C
4
, Y
4
and Bk
4
are sent to reproduction engine
112
along with the SEN signal, which determines the resolution at which the image is output.
According to application Ser. No. 08/954,226, the foregoing UCR, FILTER and SEN signals are output from black character determination unit
115
. Specifically, the UCR signal generated by black character determination unit
113
has a value from 0 to 7 indicating, from more black to less black, an amount of black component which should be removed from signals Y
1
, M
1
, and C
1
by masking UCR circuit
109
to produce signal Bk
2
. The FILTER signal produced by black character determination unit
113
is a 2-bit value in which values
0
,
1
,
2
and
3
indicated smoothing, strong edge enhancement, medium edge enhancement, and weak edge enhancement, respectively. Accordingly, the FILTER signal is input to space filter processing unit
111
to control an amount and type of filtering applied to signals Y
3
, M
3
, C
3
and Bk
3
.
The SEN signal is output from black character determination unit
113
to reproduction engine
112
, and is a 1-bit signal in which a 0 value indicates to engine
112
that printing should proceed at 200 lines per inch resolution, and the value
1
indicates that 400 lines per inch printing is required.
The values of UCR, FILTER and SEN are outputs of look-up table (LUT)
117
, which receives signals indicating a width of a character containing a subject pixel, a proximity of the subject pixel to an edge of a character, and a chromaticity of the subject pixel. Therefore, the output values of UCR, FILTER, and SEN are calculated for each subject pixel and are determined based on a detected character width, edge proximity and chromaticity corresponding to the pixel according to relationships specified by the LUT.
For example, a FILTER signal value of 1 is used for a subject pixel which is located near to an edge, has low chromaticity and is included in a relatively thin character, since such factors suggest that the pixel is within a small, black character. In another example, the SEN signal is assigned a value of 0 (corresponding to 200 lines per inch resolution) in a case that the subject pixel is not near an edge and is included in a very thick area, since larger toner dots, which provide more toner per unit area than larger dots, generate a better halftone image.
As can be seen from the foregoing, conventional image processing systems such as that described in U.S. application Ser. No. 08/954,226 “guess” the nature of a subject pixel based on several factors in order to then determine appropriate processing parameters for the pixel. One drawback to such an approach is that the guessed nature may be incorrect, resulting in inappropriate processing being performed on the pixel.
In view of the foregoing, what is needed is an image processing system to improve image processing using accurately and inexpensively detected attributes of input image data.
SUMMARY OF THE INVENTION
The present invention addresses the foregoing problems by utilizing results of block selection processing to determine processing to be applied to data of a subject pixel. Using the results of block selection processing, the nature of a subject pixel can be more specifically identified than when using the conventional systems described above, therefore more appropriate processing can be applied to the pixel data. In addition, image processing proceeds quickly using such block selection processing in combination with existing image processing hardware. Moreover, block selection processing is preferably software-based and therefore inexpensive to implement.
Particularly, the present invention is directed to an image processing system in which image data is input, block selection processing is performed on the input image data to determine types of pixel data within the image data, it is determined, based on the block selection processing, if subject pixel data represents a text pixel, and it is determined if the subject pixel data represents an edge pixel. A first processing is performed on the pixel data in a case that the pixel data is determined to represent a text pixel and an edge pixel, and a second processing is performed on the pixel data in a case that the pixel data is not determined to represent a text pixel and is not determined to represent an edge pixel.
By virtue of the foregoing features, inexpensive and appropriate image processing can be applied to image data. For example, the first processing is preferably a sharpening processing and the second processing is preferably a smoothing processing.
In a further aspect of the present invention, the inputting step includes pre-scanning image data at a first resolution, and scanning the image data at a second resolution, the second resolution higher than the first resolution, wherein the block selection processing is performed on the image data at the first resolution, and wherein the sharpening processing and the smoothing processing are performed on the image data at the second resolution. This further aspect allows overall processing to be completed quickly.
It should also be noted that the invention contemplates determining if types of pixel data are text on halftone, line art, line, title, table, halftone, frame and background. Accordingly, appropriate image processing can be applied to the image data based on these pixel data types.
This brief summa
Bayat Ali
Canon Kabushiki Kaisha
Fitzpatrick ,Cella, Harper & Scinto
Mehta Bhavesh M.
LandOfFree
Block selection-based image processing does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Block selection-based image processing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Block selection-based image processing will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3220679