JPH03157064A - Picture information compression transfer system - Google Patents

Abstract

PURPOSE:To simply and surely send a document picture with compressed data of less quantity by generating a typical character pattern form a segmented character picture and sending the generated typical pattern information and a classification code to a receiver. CONSTITUTION:A character segmentation device 12 calculates the size of an individual character pattern 31 stored in a frame memory and its existence value and sends the information to a pattern collation device 13. The pattern collation device 13 receives the calculated size of the individual character pattern 31 and its presence location information and accesses again the individual character pattern group in the frame memory and classifies them into some categories. A typical character pattern storage device 14 calculates the average pattern of each category to generate the typical character pattern and stores the typical character pattern information and the classification code in pairs and transfers them to a pattern coder 15, where they are coded.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention captures image information including printed characters recorded on a medium using a camera or scanner, and compresses the amount of information in the obtained still image. This relates to a compression and transmission method for image information that is transmitted using a computer.

[Prior Art] Conventional methods for transmitting image information of this type include the following methods. In other words, the first method is a one-dimensional encoding method that encodes and transmits the continuity of the pixel arrangement, and the second method focuses on the two-dimensional structure of pixels to compress the amount of information. Two-dimensional encoding method for transmission, and third
The fourth method is to obtain the spatial frequency components of the original image and encode and transmit only this high frequency component level.
There are two methods: a vector quantization method, and a fifth method that treats the image as a figure and extracts its geometric features and encodes them. Conventionally, these methods have been used to Information is being transmitted.

[Problem M to be solved by the invention] Among the conventional image information methods described above, the first and second methods, that is, the -dimensional encoding method and the one-dimensional encoding method, are those that have been put into practical use in facsimiles. However, it has the disadvantage that the amount of information to be transmitted is extremely large. Also,
The third frequency component transmission method had the disadvantage of being unsuitable for transmission of document images with irregular continuity, and the fourth method, the vector quantization method, Since the number of pixels in one block is taken as a fixed value, it cannot be expected to compress and transmit a document image consisting of a group of characters of arbitrary size.Furthermore, the geometrical characteristics of the fifth method The method of extracting and encoding features has the drawback that the algorithm for extracting these features is unstable, and the calculations required to extract the features are complex and enormous.

[Means for Solving the Problems] In order to solve such problems, the image information compression transmission method of the present invention cuts out individual character images from a document image containing character information, and The similarity between the images is measured to generate image information of a typical character pattern, the cutout area is classified to generate a classification code, and the generated image information of the typical character pattern and the classification code are sent to a receiving device. It is configured to be sent to.

[Operation] A typical character pattern is generated from the extracted character image, and the generated typical character pattern information and classification code are transmitted to the receiving device. As a result, the document image can be compressed and transmitted with a small amount of information, and the received document image can also be faithfully reproduced at the receiving device.

[Example] Next, the present invention will be explained with reference to the drawings.

FIG. 1 is a block diagram of a device W showing an embodiment of the image information compression and transmission system of the present invention. In the figure, 1 is a transmission system, and this transmission system 1 is configured as follows. That is, 11 is an image input device, 12 is a character area cutting device,
13 is a pattern matching device, 14 is a typical character pattern storage device, 15 is a pattern encoding device, and 16 is an information transmission device. Further, 17 is a transmission path, and 20 is a receiving system.

This receiving system 20 is configured as follows.

That is, 21 is an information receiving device, 22 is a typical character pattern storage device, 23 is an image composition device, and 24 is an image output device.

Next, FIG. 2 is an explanatory diagram showing an area where a document image is cut out. In the figure, 30 is a medium in which a document image is stored, and 31 is a cut out individual character pattern located at a position (X, Y) on the medium 30 and having a size (W, H).

Generally, printed document images have many regularities and characteristics, but there is no conventional image information transmission method that focuses on and utilizes these regularities and characteristics of document images. The image information compression and transmission system of the present invention focuses on and utilizes the two-dimensional structure of a document image and the similarity between individual characters. Here, the similarity between individual characters means that each individual character pattern 31H has some correlation, and means the redundancy of the original image.The image information compression and transmission method of the present invention eliminates this redundancy. By deleting the information, the amount of information is reduced and transmitted.

Furthermore, regarding the encoding method of the character area, there are the following three methods. That is, the first method treats character areas in the same way as image areas, but adapts its parameters to the characters. The second method is to cut out individual characters and represent them with a typical character pattern. Furthermore, the third method is to recognize characters and treat them as character codes. The first method also includes a method that utilizes the two-dimensional structure of a figure. In the second method, the image is accurately reproduced because the character pattern is retained. In addition, in the third method, although the image information is lost, it is compressed into character information, so a high compression rate can be achieved, and the information processing performed in the next stage is facilitated. 9 Image information compression of the present invention The transmission method is the second method of the three methods described above.

Next, the operation of the image information compression transmission system according to the present invention will be explained in detail.

The image input device 11 uses a medium 30 on which a printed character string is recorded.
The above optical density information is photographed by a photoelectric conversion element, transferred to a frame memory (not shown), and stored. Further, the character cutting device 12 calculates the size and location of the individual character pattern 31 stored in the frame memory and transmits this information to the pattern matching device 13. The location, size, and classification code corresponding to the cutout region number are shown.

Then, the pattern matching device 13 receives the calculated size and location information of the individual character pattern 31, accesses the individual character pattern group in the frame memory again, and classifies them into several categories. do. That is, for example, the individual character pattern of cutout area number 1 is classified into the classification code A, and the cutout area 2 is classified into the classification code B, and this information is sent to the typical character pattern storage device 14. The typical character pattern storage device 14 calculates the average pattern of each category, generates this average pattern, that is, a typical character pattern, and stores this typical character pattern information, the category number,
That is, they are stored in pairs with classification codes, and are transferred to the pattern encoding device 15 for encoding.

The typical character pattern information and classification code encoded by the pattern encoding device 15 are transmitted to the information receiving device 21 via the information transmitting device 16. Note that the information transmitting device 16 and the information receiving device 21 are modulators and demodulators provided to transmit encoded information via the transmission path 17.

The typical character pattern storage device 22 receives the typical character pattern and classification code received and demodulated by the information receiving device 21, reproduces it as a two-dimensional image, and transfers it to the image synthesis device 23 to synthesize it. 24 outputs the reproduced image synthesized by the image synthesis bag W23.

As explained above, according to the image information compression transmission method of the present invention, when a general document image in which an arbitrary character string is recorded on the medium 30 is targeted, not only the character information in the document but also the Even the unique font information can be compressed and transmitted. Therefore, the receiving system 20 can faithfully reproduce the image information contained in the original image. Further, in order to synthesize images using an ideal typical character pattern, physical noise in the original image, spatial quantization error of the image input device 2, etc. are removed.

[Effects of the Invention] As explained above, the image information compression transmission method of the present invention has the following advantages:
Generate typical character patterns from the extracted character images,
Since the generated typical character pattern information and classification code are configured to be sent to the receiving device, document images can be transmitted easily and reliably with a small amount of compressed information, and the received document images can be faithfully reproduced. There is an effect that it can be done.

[Brief explanation of the drawing]

FIG. 1 is an apparatus configuration diagram showing an embodiment of the image information compression and transmission method of the present invention, and FIG. 2 is an explanatory diagram illustrating the arrangement of individual characters recorded on a medium. 11... Image input device, 12... Character area cutting device, 13... Pattern matching device, 14.22
... Typical character pattern storage device, 15 ... Pattern encoding device, 16 ... Information transmission device, 17 ...
Transmission path, 21...information receiving device, 23...image composition device, 24...image output device, 30...medium, 31...individual character pattern.

Claims (1)

[Claims] Individual character images are extracted from a document image containing character information, and the similarity between the extracted partial images is measured to generate image information of a typical character pattern.The extracted regions are also classified and a classification code is generated. 1. An image information compression and transmission method, characterized in that image information can be compressed and transmitted by generating image information of a typical character pattern and transmitting the generated image information of a typical character pattern and a classification code to a receiving device.