Image analysis – Pattern recognition – Classification
Reexamination Certificate
1998-02-06
2001-08-07
Boudreau, Leo (Department: 2721)
Image analysis
Pattern recognition
Classification
C382S176000, C382S218000, C382S266000, C358S462000, C358S461000
Reexamination Certificate
active
06272249
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to an image processing device for determining an image area of image data obtained from an original by scanning, which is usable for digital copying machines, scanners and so on.
In conventional digital copying machines and scanners, image data inputted from an original through a CCD (Charge-Coupled Device) sensor or the like is subjected to processing for improving the image quality. For example, Japanese Laid-open Patent Publication (TOKKAI HEI) No. 4-188948 discloses an image processing device that discriminates a particular area of the image based on a feature extracted read-in image data and executes processing for further improving the quality of the image. According to this method, an average dispersion of differences of image data is determined relative to an average value obtained in a particular area of the image to discriminate the particular area as a character area if the average dispersion value is larger than the average value or as a halftone area if the average dispersion value is smaller than the average value, and the image processing is made adaptively to the discrimination result.
Japanese Laid-open Patent Publication (TOKKAI HEI) No. 4-270561 discloses another image processing device that sets two upper and lower thresholds Tb and Tc (Tb >Tc), discriminates a particular area as a character area if the pixel density of the area is larger than the threshold Tb and smaller than the threshold Tc or as a halftone area if the pixel density is between Tb and Tc, and performs the image processing according to the discrimination result.
The former of the above-mentioned two image-discriminating methods involves a problem that the difference between an average dispersion value and an average value does not always sufficiently represent the feature of an image area and, therefore, may sometime make an error in discrimination of the image. This method may sometime have inadequate criteria and unsuitable average values as thresholds. Namely, further improvement of area discrimination accuracy is required.
The latter image discriminating method may reflect only a limited feature in the discrimination result since it discriminates an image area according to only two thresholds. Consequently, it cannot achieve a high accuracy discrimination of the image and have a high probability of misjudgment of the image, making it difficult to improve the image quality by further image processing.
Accordingly, the present applicant has proposed an image discriminating method which, as disclosed in Japanese Patent Application No. 8-238205, is free from the above-mentioned problems. Namely, this method estimates a feature value of a particular image area according to differences between optical densities of pixels neighboring to each other in horizontal and vertical directions in the area and discriminates the classification of that particular image area according to the feature value.
Therefore, the present invention has as its primary object the provision of an image processing device which can attain a further improved accuracy of discrimination of the classification of a particular image area to be processed for improving its quality.
SUMMARY OF THE INVENTION
To achieve the above-mentioned object, the present invention provides an image processing device for reading image information, discriminating the classification of a particular image area composed of a plurality of pixels and processing the image area by filtering according to the classification resulted from the discriminating, which comprises feature calculating means for determining a feature value of the particular area according to differences in optical densities of pixels disposed in vertical, horizontal, main diagonal and sub diagonal (perpendicular to the main diagonal) directions within the particular area, and area discriminating means for determining the classification of the particular area on the basis of the estimated feature value.
In this device, the classification of a particular area of an image can be discriminated at increased accuracy since a feature value used for discriminating the classification of a particular image area is determined according to differences between optical densities of pixels disposed in four (horizontal, vertical, main diagonal and sub diagonal) directions in the area. The image area may be classified, by way of example, into a photographic area, a character area and a halftone dot area.
The feature value estimating means comprising: least value obtaining means for calculating a total value of differential densities of pixels neighboring to each other in the vertical direction in a particular area, a total value of differential densities of pixels neighboring to each other in the horizontal direction in the particular area, a total value of differential densities of pixels neighboring to each other in the main diagonal directions in the particular area and a total value of differential densities of pixels neighboring to each other in the sub diagonal directions in the particular area and determining a least one of the total values of differential densities as a feature value of the particular area, and differential value obtaining means for determining a difference between the maximal density value and minimal density value of the pixels within the particular area as a feature value of the particular area.
The feature value estimating means may also be composed of least value obtaining means for adding a total value of differential densities of pixels neighboring to each other in the vertical directions in the particular area to a total value of differential densities of pixels neighboring to each other in the horizontal directions in the particular area, adding a total value of differential densities of pixels neighboring to each other in the main diagonal directions in the particular area to a total value of differential densities of pixels neighboring to each other in the sub diagonal directions in the particular area and determining a least one of two summed values as a feature value of the particular area and differential value obtaining means for determining a difference between the maximal density value and minimal density value of the pixels within the particular area as a feature value of the particular area.
The feature amount of the particular image area, which was determined on the basis of the differential densities of pixels disposed in the horizontal and vertical directions and/or diagonal directions, can indicate the degree of complexity of the particular area and can also reflect the correlation between pixels on the classification of the area. This improves the accuracy of discrimination of the image area classification.
REFERENCES:
patent: 5436644 (1995-07-01), Motoi et al.
patent: 5530559 (1996-06-01), Sato
patent: 5629752 (1997-05-01), Kinjo
patent: 5902987 (1999-05-01), Coffman et al.
patent: 0 663758 A2 (1995-01-01), None
patent: 0 710004 A2 (1995-10-01), None
patent: 4-188948 (1992-07-01), None
patent: 4-270561 (1992-09-01), None
patent: 10-084475 (1998-03-01), None
Kobayashi Setsuya
Murakami Takeshi
Boudreau Leo
Desire Gregory
Sharp Kabushiki Kaisha
LandOfFree
Image processing device 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, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing device will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2526769