Image processing using vector data to reduce noise

Details Image processing using vector data to reduce noise Image processing using vector data to reduce noise

1994-07-21

2001-02-06

Mehta, Bhavesh (Department: 2721)

Image analysis

Image enhancement or restoration

Edge or contour enhancement

C358S451000

Reexamination Certificate

active

06185341

ABSTRACT:

BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to smoothing processing of outline-vector data or magnification-varying processing of a digital binary image, and more particularly, to an imager processing method and apparatus for obtaining high-quality outline data or a high-quality image subjected to magnification-varying processing using contour information.
2. Description of the Related Art
The above-described technique has been used as a technique for producing digital fonts (a typeface) used in DTP (desktop publishing) apparatuses, word processors and the like.
The technique is used for automatically extracting rough contour data in a system of editing and forming outline fonts. That is, predrawn large-format (analog) characters are read by a digital scanner, and rough contour data are automatically extracted based on bit-map data of digitized large-format characters, each of which comprises at least about several hundreds of pixels (picture elements)×several hundreds of pixels. A font designer forms final outline characters by performing correction while comparing bit-map data of such large-format characters of an original with rough contour data on a display.
On the other hand, a technique has, for example, been proposed in which high-quality vectorized fonts are completely automatically formed at high speed from dot characters, each of which comprises rather small dots of, for example, about 48 dots×48 dots (Murayama and Watanabe, “Automatically Improved Vectorized Fonts”, Preprint of National Meeting of Gazo Denshi Gakkai (The Japan Society of Image Electronics) 27, pp. 117-120 (1988)).
An SPC (selective processing conversion) method (see Matsumoto and Kobayashi, “A Study of Evaluation of Picture Quality in Resolution Conversion in a Facsimile”, Gazo Deri shi Gakkai Shi (Journal of Japan Society of Image Electronics), vol. 12, no. 5, pp. 354-362 (1983)), in which each pixel in an original image is repeatedly used a number of times equal to the desired magnification or is periodically skipped in accordance with varying magnification, has generally been used for magnification-varying processing of a digital binary image. However, since this method has problems in picture quality such that, for example, oblique-line portions have a stepwise shape, various other methods have been proposed, such as a smoothing method which works by referring to a pattern of pixels surrounding a target pixel (Imanaka, Semasa and Ono, “Provision of High Picture Quality by Smoothing Processing of a Received Image in a Facsimile”, Preprint of Annual Meeting of Gazo Denshi Gakkai (Japan Society of Image Electronics), no. 18 (1991)), a projection method in which an original image is projected onto a conversion image surface having different line density, and the values of pixels of the converted image are determined by performing binary-coding processing of integrated values relating to respective pixels within this; surface using threshold logic (Arai and Yasuda, “A Study of Line-Density Conversion in a Facsimile”, Gazo Denshi Gakkai, Shi (Journal of Japan Society of Image Electronics), vol. 7,, no. 1, pp. 11-18 (1978)), and the like.
As described above, in the field of DTP, a method has generally been used in which digital fonts (a typeface) are stored in coordinate representation, and are subjected to dot development after changing magnification of stored coordinate values in accordance with desired magnification.
Problems in the above-described known technique will now be considered.
The outline-vector smoothing method in the system of editing and forming outline fonts is a method of roughly reducing the size of a character in which vectorization is performed based on the diagram of the character comprising a very large number of dots. Hence, contours of lines having a one-dot width or a two-dot width, and of fine points or holes comprising a small number of dots, i.e., about a few dots, will not be extracted or will be smoothed over. Therefore, such a method is not suitable for forming outline-vector data of fine-line images or complicated images, from the viewpoint of accuracy and picture quality. Furthermore, a considerable amount of calculation is required for smoothing processing, and therefore real-time calculation cannot be expected if an inexpensive computer system is used.
In the method of automatically forming high-quality vector fonts from dot characters comprising a small number of dots with high speed, beautifully shaped dot fonts are assumed as input fonts. Therefore, consideration is not taken for processing for noise, such as notches, isolated points and the like, which will be produced in an image read by a scanner or the like. The method also has the problem that smoothing processing is performed even on fine information having a one-dot width, and therefore fine features in the image are lost.
The smoothing method which works by referring to a pattern of pixels surrounding a target pixel has the problem that the target pixel can be processed only by a fixed magnification of two or four times in the main-scanning direction and the sub-scanning direction. The projection method has the problem that effective magnification is limited to ½ to two times, and high picture quality cannot be expected when the range of magnification is expanded to other values.
The method of obtaining character diagrams subjected to magnification-varying processing by extracting outline vectors from bit-map fonts uses only bit-map fonts as input data, and assumes that information relating to the properties of an input image, such as typeface information (Ming-dynasty characters, Gothic characters, katakana, hiragana or the like) or the like, is provided as additional information. Furthermore, this method cannot deal with noise, such as notches, isolated points and the like, contained in an input image, and fine information having a one-pixel width. Renee, it is not suitable to apply this method to a general binary image.
SUMMARY OF THE INVENTION
The present invention provides an image processing method which can be applied even to outline vectors obtained from an image having a complicated structure and noise, which is produced by reading, for example, a line drawing such as a weather chart, a business document, or the like, by an image reader or the like with a relatively small amount of calculation, by determining a target outline vector after smoothing processing in accordance with a pattern of a target outline vector and outline vectors near the target outline vector, and introducing a rule for determining whether a fine structure having a one-dot width must be removed as noise, must be preserved as a salient point without being subjected to smoothing processing, or must be made to be a point on a smooth contour by performing smoothing processing.
It is an object of the present invention to obtain a digital binary image having high picture quality subjected to magnification-varying processing with a desired (arbitrary) magnification, by extracting outline vectors from a binary image, forming outline vectors subjected to smooth magnification-varing processing with the desired (arbitrary) magnification in a state of representation of the extracted outline vectors, and restructuring a binary image from the outline vectors subjected to the smooth magnification-varying processing.


REFERENCES:
patent: 3755780 (1973-08-01), Sammon et al.
patent: 4153896 (1979-05-01), White
patent: 4162482 (1979-07-01), Su
patent: 4648120 (1987-03-01), Chittineni
patent: 4777651 (1988-10-01), McCann et al.
patent: 4949281 (1990-08-01), Hillenbrand et al.
patent: 4979229 (1990-12-01), Noolenaar
patent: 5093870 (1992-03-01), Watanabe
patent: 5222154 (1993-06-01), Graham et al.
patent: 5644366 (1997-07-01), Ushida et al.

Affiliated with

Also associated with

Rating

Image processing using vector data to reduce noise 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 using vector data to reduce noise, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing using vector data to reduce noise will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2565749