Browse Prior Art Database

Direct Generation of Compressed Images of Text String Data

IP.com Disclosure Number: IPCOM000085721D
Original Publication Date: 1976-May-01
Included in the Prior Art Database: 2005-Mar-02
Document File: 5 page(s) / 95K

Publishing Venue

IBM

Related People

Chesarek, DJ: AUTHOR

Abstract

This generation program will directly generate the compressed representation of a picture element (PEL) matrix image of text material. It significantly reduces the number of instruction executions required to generate the compressed representation of the PEL matrix image of text material.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 47% of the total text.

Page 1 of 5

Direct Generation of Compressed Images of Text String Data

This generation program will directly generate the compressed representation of a picture element (PEL) matrix image of text material. It significantly reduces the number of instruction executions required to generate the compressed representation of the PEL matrix image of text material.

The input to the generation program is a string of EBCDIC text and an entry point to a symbol dictionary, which contains the PEL matrix representation of each possible symbol. The output is a compressed representation of the image.

The generation program is based on two interdependent characteristics of the predictor used in the decompressor and of the character design. Fig. 1 shows the area involved in the prediction process. Basically, the prediction process uses only bits which are one cell to the left of or one cell above the cell to be predicted.

Most character designs contain a "white" border to provide visual isolation between adjacent characters.

The generation program to be described herein requires at least a PEL wide white border on the right edge and bottom of each character cell. The changes required for any other edge and end border pair can be achieved by a simple rearrangement of the program logic which links character images together. This linkage can also be designed to accommodate variable spacing between nominal character cells.

The combined effect of the limited amount of prediction used in the decompressor and the one cell which border, is to remove any adjacent character cell prediction dependencies. The lack of intercharacter cell prediction dependencies means that it is possible to generate the compressed representation of a page of text directly from the text; the more conventional approach being to generate an uncompressed PEL matrix for the entire page, and then to program convert the PEL matrix data from uncompressed format to compressed format. The direct generation compressed representation approach requires significantly less CPU resource than the conventional approach, due to the significant reduction in the amount of bit string shifting required to develop the uncompressed PEL matrix.

A Symbol Dictionary is provided which contains one entry point for each character to be generated. All symbols are assumed to be the same number of raster lines in height, but may have varying widths (i.e., proportional spaced characters). Since the program is working with a compressed representation of character data, it does not need to know the PEL width of the character.

There is no intercharacter prediction dependency, therefore, the data representing the character image pattern can be stored in compressed form. There will be one compressed string for each scan line in the character cell. Since the string can be of varying length, each string has a length attribute which is used by the image generation program.

1

Page 2 of 5

In addition to the compressed string data, ea...