JPS604373A - Pattern information compressing system - Google Patents

Abstract

PURPOSE:To improve a pattern compressing ratio to about 1/100 so that a two- dimensional pattern can be compressed, by encoding the recognized result of the two-dimensional pattern after positional information is added to the result. CONSTITUTION:Information of a document 1 is converted into an electric signal by means of a scanner 2 and the electric signal is sent to a cutting section 3. The cutting section 3 detects character parts from the electric signal and sends the character parts to a character pattern recognizing section 4 and, at the same time, supplies positional information of the stubbed characters, such as position, size, etc., to an encoding section 5. Moreover, a recognized result code from the recognizing section 4 is also sent to the encoding section 5 where the recognised result code is encoded and code information of one character is obtained. The code information is given to a storage device 6. By having a character code together with information of a two-dimensional plane, a two-dimensional pattern can be reproduced in the form of a fair copy.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a pattern compression method, and more particularly to a pattern compression method suitable for compressing information on documents created by a word processor.

[Background of the invention]

One of the conventional methods for compressing patterns is run-length encoding. This method sequentially scans a two-dimensional pattern in one dimension and encodes the continuous length of the pattern, and is widely used because it is relatively simple. But there was a vanishing point.

[Purpose of the invention]

The purpose of the present invention was to eliminate the above-mentioned drawbacks, and it is an object of the present invention to provide a pattern compression method capable of compressing a two-dimensional pattern by increasing the compression rate to approximately 2.

[Summary of the invention]

To achieve this objective, the present invention is characterized in that the two-dimensional pattern recognition results and position information are encoded to perform information compression.

[Embodiments of the invention]

An embodiment of the present invention will be explained below with reference to FIG.

FIG. 1 is an example of a document created using a word processor or the like. The present invention performs information compression on such documents using the following procedure. In other words, characters are cut out, all of them are recognized, and then coded. Furthermore, information such as the position information of the cut characters and the size of the characters is added to the recognition result to create compressed information for one character. Do this over the entire document.

FIG. 2 shows an embodiment of the present invention. In FIG. 2, information on a document 1 is converted into an electrical signal by a scanner 2. The extraction unit 3 detects all character parts and recognizes the character pattern by the recognition unit 4.
At the same time, the position and size of the cut out characters are detected. The recognition result code of the recognition unit 4 and character information (position, size, etc.) on the two-dimensional plane detected by the cutting unit 3 are all encoded in the encoding unit 5 to obtain code information for one character, and this code is converted into, for example, It is stored in the storage device 6.

By having the character code and information on the two-dimensional plane in this way, it is possible to print and reproduce the two-dimensional pattern. That is, the original sentence v can be reproduced by restoring the character code and size information to a character pattern using a character pattern generator, and determining the position on the two-dimensional plane from the character position information.

Next, specific processing will be explained using FIG. 1 as an example. Cut out the character "sentence" in the character string "document 1" currently written on document 11 and find the position information (for example, the position of the center of gravity) xl
, yt and size information (for example, double 9 characters) M are obtained, xIylM is sent to the encoding unit. Further, from 4, the code CB of "1 sentence" is sent to 5, fl, and the output of 5 produces a code output as shown in FIG. 3, for example.

If the position information of the character is, for example, the center of gravity, the position of the center of gravity will differ depending on the character type, so if C is left as is, the reproduced character string will fluctuate. Therefore, it is necessary to normalize xt and yt by character pitch and line pitch before reproduction.

FIG. 4 is an explanatory diagram for normalization. In FIG. 4, Xl is the center of the character string, and ΔX++ ΔX2 is the deviation of the double knot b from Xl. Now, X, =X, +ΔXl, x24
Since X1- is 1, the position of the character in the line direction can be set to +.

〔Effect of the invention〕

According to the present invention, the compression ratio can be improved by about one order of magnitude compared to conventional compression methods such as run length. For example - font size 1
When expressed as a 64x64 mesh, it is about . However, in the present invention, when one sentence is 2000 x 2000 meshes and 12000 character types, 1 in the X direction for position information.
1 bit, 11 bits in the Y direction, and 11 bits for the character type, a total of 33 bits or less information.

Furthermore, since the information is stored as a character code, there is no deterioration during playback, and documents can be easily searched by using the corresponding code.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram for understanding the present invention, FIG. 2 is a configuration diagram of an embodiment of the present invention, and FIGS. 3 and 4 are diagrams for explaining the present invention.

Claims (1)

[Claims] 1. A pattern information compression method characterized in that a two-dimensional pattern is recognized, and the information pressure mk of a fly-dimensional pattern is obtained by adding plane information of the two-dimensional pattern to the recognition result code and encoding it.