JPH08102853A - High-speed character determination method in image processing - Google Patents

Abstract

(57) [Abstract] [Purpose] The present invention provides a device for storing documents, etc., which occur daily in a large-capacity storage medium to improve the efficiency of document search, etc.
It is intended to shorten the image processing time when it is necessary to perform image processing such as image enlargement and reduction. [Structure] When performing image processing, before determining the character area and the photo area of the image, a process for roughly and quickly determining whether the image is a full-face character image is inserted, and the image is determined to be a full-face character image. If so, the time-consuming detailed character area and photograph area determination processing is not performed. According to the present invention, when performing image processing such as image enlargement and reduction, it is possible to shorten the image processing time particularly for a full-face character image.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention requires different processing for each area when performing image processing such as enlargement and reduction on image data having different areas such as characters and photographs. For example, the present invention relates to an area determination method when it is necessary to determine each area.

[0002]

2. Description of the Related Art When displaying image data stored in a storage medium on a display device such as a display smaller than the image size, when performing reduction printing from a printing device, or when performing image editing processing on the image data, Data expansion and contraction processing is required. When the image data includes different areas such as characters and photographs, if the same enlargement / reduction processing method is applied to the different areas, the image quality of one image area may be lower than that of the original image. Such a phenomenon may occur, for example, when there is a character area and a photo area and the same reduction processing is performed on both areas, the image quality of the photo area deteriorates with respect to the original image. In such a case, it is necessary to perform different image processing on both areas. For this purpose, determination processing for each area is required as a pre-stage for performing different image processing for each area. Usually, this area determination method divides an image into areas of a certain size (for convenience, this area is hereinafter referred to as a window), and determines from a pixel pattern in each window whether it is a character area or a photograph area. If the size of this window is made smaller, finer area determination is possible, but the smaller the amount of pixel information, the more difficult it becomes to determine characters and photographs. Therefore, the actual size of the window needs to be set appropriately.

Conventionally, after this area determination is performed on the entire surface of the image data, image processing is performed on each area to obtain an output image.

[0004]

As in the prior art, after performing the area determination of the character area and the photograph area on the entire surface of the image,
The method to execute the processing suitable for each area is
It is considered to be an appropriate method for image data that always has a photographic area. However, when storing a document or the like that occurs daily in a storage medium, most of it is an image of characters only, not including a photograph. When performing image processing on this character-only image data, it is possible to determine the area between the character portion and the photograph portion even though the image data consists of only the character portion that can be processed at a relatively high speed compared to the photograph portion. It takes much time, and high-speed image processing cannot be performed. The time required for this region determination may become longer than the time required for image processing of the character region when the entire region of the image is examined, and the region determination time becomes a large overhead.

According to the present invention, by shortening the area determination time for an image consisting of only characters that occur frequently in everyday life,
The purpose is to keep the image processing time low.

[0006]

[Means for Solving the Problems] In order to reduce the image processing time in an image containing only characters, it is quickly determined whether the image includes only characters or includes a photograph,
In the case of only characters, it is necessary to perform character area processing immediately. Therefore, in the present invention, before the detailed area determination is performed and the character area and the photo area are separated, the image is divided into windows that are relatively large compared to the size of the window used when performing the detailed area determination, and all the windows are divided. On the other hand, a process for quickly determining whether or not the image includes a photograph is provided by not examining the character region or the photograph region but examining only a part of the windows. In this process, when all the investigated windows are judged to be character areas, they are regarded as full-scale character images. Further, not all of the examined windows are determined to be character areas, and even if some windows are determined to be photo areas, if the number is a certain number or less, it is considered as a full-face character image. The number used for this determination and the number of windows to be investigated must be determined so as not to impair the image quality after image processing. Further, it is necessary to set the processing time for determining whether the image is a rough full-face character image to be sufficiently shorter than the ordinary detailed determination time for a character area or a photograph area. When it is determined in this determination process that the image is a full-face character image, the character image process is performed on the entire image. On the other hand, when it is determined that the image has a photographic region, the conventional detailed region determination is performed, and image processing corresponding to each region is performed.

As described above, by roughly determining whether the image is a full-face character image before making a detailed region determination, it is expected that the image processing speed for an image of only characters that frequently occur everyday will be improved.

[0008]

By performing the above-described rough full-face character image determination processing, it is expected that the image processing speed for the full-face character image will be improved. However, although this method can perform image processing faster in the case of full-face character images than in the conventional method, in the case of an image with a photo area, it is extra due to the determination process of whether it is a rough full-face character image. It takes a long time to process the image. However, in addition to a full-face character image that often occurs in everyday life, even if the image includes a photo area, the image processing of the normal photo area takes a longer processing time than the image processing of the character area. The high-speed processing for determining whether the image is a full-face character image introduced in 2. does not cause much overhead.

As described above, according to the present invention, in an apparatus for storing documents that occur daily in a large-capacity storage medium to improve the efficiency of document retrieval and the like, when it is necessary to perform image processing such as image enlargement and reduction. In particular, this is a processing method that is effective in reducing the image processing time of a full-face character image.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to FIGS. FIG. 1 shows the image processing method used in the present invention. In FIG. 1, an input image 1 is an original image that undergoes image processing such as enlargement and reduction, a full-face character image determination process 2 is a processing module that determines whether the input image is a full-face character image, and a region determination 3 is a character region of the input image. Is a processing module for determining a photographic area, photographic processing 4 is an image processing module in a photographic area, character processing 5 is an image processing module in a character area, and output image 6 is an output image after completion of image processing processing. In the image processing method of FIG. 1, when a full-face character image is input to the input image 1, an output image in which the character process 5 is executed immediately after passing through the full-face character image determination process 2 is obtained. That is, the processing is performed along the path indicated by the solid arrow. here,
The processing time of the full-face character image determination processing 2 is set to be shorter than the processing time of the area determination 3. As a result, when the input image is a full-face character image, the processing time can be shortened as compared with the conventional processing that passes the area determination 3. Also,
When an image having a photograph area in the input image 1 is input, the image is processed by the route indicated by the dotted arrow.

FIG. 2 shows an example of the entire character area determination processing 2 of FIG. In the processing here, first, the input image is divided into areas of about 1 cm square, which is larger than the area division performed in the detailed area determination of the area determination 3 of FIG. Then, instead of judging whether all windows are character areas or photograph areas, every other window is checked, such as judgment area 1, judgment area 2, judgment area 3 ... It can be determined at high speed whether it is a character image.

By performing this full-face character area determination processing, it becomes possible to perform high-speed processing such as image enlargement and reduction processing in a full-face character image.

[0013]

As described above, according to the present invention, in an apparatus for storing documents that occur daily in a large-capacity storage medium to improve the efficiency of document retrieval, when it is necessary to perform image processing such as image enlargement and reduction, In particular, it is possible to shorten the image processing time of the full-face character image.

[Brief description of drawings]

FIG. 1 is a diagram for explaining an image processing method used in the present invention.

FIG. 2 is a diagram showing a processing example of full-face character area determination processing 2 in FIG.

[Explanation of symbols]

Region 1 ... Region n + 2: Larger divided regions than in the case of performing the detailed region determination of the region determination 3 and set every other region.

Claims (1)

[Claims] 1. A means for reading a document containing characters, photographs, etc. as image data, a means for reading / writing the read image data from / to a non-volatile storage medium such as an optical disk, and an enlargement of the image data stored in the storage medium. In order to reduce the image processing time for an image of only characters that often occur in everyday life in a device that has a function of performing image processing such as reduction and rotation, before performing the area determination of the character portion and the photograph portion of the image data, An image processing method that reduces the area determination time overhead in the image processing time of a full-face character image by decimating appropriately for all image regions and determining at high speed whether or not it is a full-face character image by appropriately thinning out.
Alternatively, a determination method for determining images at appropriate intervals to quickly determine whether the image data is a full-face character image.