Reformatting binary image data to generate smaller compressed image data size

Abstract

In various exemplary embodiments, systems include a segmentor to segment text binary image data to a first plane. A subtractor subtracts text binary image from binary image data to generate a non-text binary image data in a second plane. A converter converts non-text binary image data in the second plane into non-text gray scale image data in the second plane. A first compressor compresses the text gray scale image data in the first plane. A second compressor compresses the non-text gray scale mage in the second plane.

Description

BACKGROUND OF THE INVENTION [0001] 1. Field of Invention [0002] This invention is directed to reformatting binary image data. [0003] 2. Related Art [0004] Binary image data means that each pixel in the image data is a discrete value “1” or “0” representing either a black pixel or a white pixel. Documents to be converted into binary image data are typically continuous tone images. [0005] Continuous tone image means that the image is represented by various shades of color or various shades of gray. Various processes can be used to convert continuous tone image into binary image data, and yet give an impression of continuous tone image. [0006] One such process, the halftoning process uses a mathematically stored screen pattern. The process systematically compares each pixel's continuous tone value with the value of the screen. If the continuous tone value of the pixel is less dense than the screen value, a white pixel is produced. On the other hand, if the pixel value is more dense tha...

AI Technical Summary

Benefits of technology

[0012] To accommodate for image data having different compression needs, a scheme such as the mixed raster content (MRC) format may be used. FIG. 1 illustrates one exemplary embodiment of an MRC format. A document 100 is segmented into a selector plane 110 (110′) and a background plane 120 (120′). Although two planes are shown in the figure, it could be N planes depending on the application. The selector plane 110 (110′) contains high spatial frequency image data. The background plane 120 (120′) contains smoothly varying low spatial frequency image data. Segmenting the document 100 into the selector plane 110 (110′) and the background plane 120 (120′) provides for different compression schemes to be applied to different planes. Thus, the most appropriate compression scheme for each image data may be applied.

[0013] This invention provides systems and methods for reformatting binary image data into two or more planes.

[0014] This invention provides systems and methods for segmenting binary image data into text binary image data and non-text binary image data and creating two or more planes from the segmented binary image data.

[0015] This invention provides systems and methods for converting binary image data into gray scale image data and segmenting the gray image data into two or more planes.

[0016] This invention provides systems and methods for enhancing binary image data and / or gray scale image data during segmentation into two or more planes.

[0017] In various exemplary embodiments, the systems include a converter to convert binary image data into gray scale image data. A segmentor segments gray scale image data into high spatial frequency image data in a first plane and low spatial frequency image data in a second plane. A first compressor compresses the high spatial frequency image data in the first plane and a second compressor compresses the low spatial frequency image data in the second plane.

Problems solved by technology

Binary image data of converted documents occupy a very large memory space in an image data processing system.

Large binary image data also consumes large bandwidth when transmitted, for example, over a local or wide area network, over an intranet, an extranet or the Internet, or other distributed networks.

However, binary image data that are error-diffused do not compress well using CCITT G3 / G4.

Additionally, binary image data containing both high spatial frequency data and smoothly varying low spatial frequency data may not optimally compress well together.

Images