JPH08223409A - Image processing unit and its method - Google Patents

Abstract

PURPOSE: To realize an image processing unit and its method in which a kind of a received image is discriminated. CONSTITUTION: An image processing unit 102 discriminates the kind of an image received from an image input device 101 and when the result of discrimination indicates an' image comprising a single density object, simple thresholding is applied to the input image and when not pseudo intermediate tone thresholding is applied to the received image and the processed image is filed and registered in a storage device 103. Thus, when lots of images are filed, since the kind of the image is automatically discriminated, it is not required for the user to confirm the kind of each image.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus and a method thereof, for example, an image processing apparatus for discriminating the type of an input image.

[0002]

2. Description of the Related Art An electronic filing system is known in which an image input by a scanner is subjected to image processing and stored as electronic information. This electronic filing system, regardless of the type of image, binarization to save the image,
This is to perform processing such as image area separation, OCR (optical character recognition), and filing registration on the input image. Here, image area separation is a process of dividing an input image into regions such as characters, photographs, graphs, and tables, and OCR is a process of recognizing a character image and converting it into a character code.

[0003]

However, the above-mentioned conventional example has the following problems. That is, since the same process is performed on the input image regardless of the type of image, the process may not be appropriate depending on the type of image. For example, when inputting a document containing a photographic image, it is desirable to binarize while leaving the tonality of the photo, but the tone cannot be reproduced in a simple binarized image. There is a problem that parts are crushed.

There is also an electronic filing system capable of switching processing depending on the type of image, but this switches processing based on the user's instruction, that is, based on the image type judged by the user. The system itself does not judge the type of image and switch the processing. Therefore, the user needs to confirm the type of each image, and when filing a large number of images, a troublesome work is required.

An object of the present invention is to solve the above problems, and an object of the present invention is to provide an image processing apparatus and method capable of discriminating the type of an input image.

[0006]

[Means for Solving the Problems] and

The present invention has the following structure as one means for achieving the above object.

The image processing apparatus according to the present invention calculates the luminance distribution from the input image, divides the luminance distribution obtained in the calculating step into a plurality of luminance sections, and divides in the dividing step. A determination step of determining the type of the input image based on the bias of each luminance distribution.

Further, a first calculation step for obtaining the entire luminance distribution from the input image, a division step for dividing the entire luminance distribution obtained in the first calculation step into a plurality of luminance sections, and a division for the division step. The second calculation step of calculating the value representing the bias of each of the brightness distribution, the judgment step of judging the shape of the overall brightness distribution based on the calculation result of the second calculation step, and the judgment step A determination step of determining the type of the input image from the shape of the overall luminance distribution.

Further, the image processing method according to the present invention comprises a calculating means for obtaining the luminance distribution of the input image, a dividing means for dividing the luminance distribution obtained by the calculating means into a plurality of luminance sections, and the dividing means. And a discriminating means for discriminating the type of the input image based on the bias of each of the luminance distributions divided by.

Further, the first calculating means for obtaining the overall luminance distribution from the input image, the dividing means for dividing the entire luminance distribution obtained by the first calculating means into a plurality of luminance sections, and the dividing means. Second calculating means for calculating a value representing the bias of each of the divided luminance distributions, judging means for judging the shape of the overall luminance distribution based on the calculation result of the second calculating means, and the judging means And a determining unit that determines the type of the input image based on the determined shape of the entire luminance distribution.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An image processing apparatus according to an embodiment of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing the schematic arrangement of an electronic filing system including an image processing apparatus according to an embodiment of the present invention.

In FIG. 1, reference numeral 101 denotes an image input device, which is composed of, for example, an image scanner or the like, and is for reading an image of a document to be electronically filed. Reference numeral 102 denotes an image processing apparatus according to the present exemplary embodiment, which performs processing described below on an image input from the image scanner 101. The storage device 103 includes a hard disk, an optical disk, or a magneto-optical disk as a storage medium, and the image processing device 102
The image output from is stored. An image output device 104 is composed of, for example, a monitor such as a CRT or LCD, or a printer, and forms an image based on the image information output from the image processing device 102.

The image processing device 102 controls the operations of the image input device 101, the storage device 103, and the image output device 104, controls the entire electronic filing system, and has a function of searching and outputting electronic filing and a filed image. I have it. The image processing apparatus 102 can be configured as dedicated hardware, but can also be realized by, for example, a combination of a personal computer and software for an electronic filing system.
The image to be electronically filed is not limited to the one formed on the recording paper or the like, and may be an image formed from document data or image data created by word processing software or DTP software.

FIG. 2 is a flowchart showing an example of a procedure of electronic filing executed by the image processing apparatus 102.

First, in step S501, image data is input from the image input device 101. Here, the input image data is, for example, 8-bit multivalued data (8-bit gray scale), and the value 0 is black and 255 is white.

Subsequently, image discrimination is performed in step S502.
Specifically, documents and drawings that consist of simple images (hereinafter referred to as "single density target image"), documents in which complicated images such as photographs and drawings are placed, and density changes in the background A document and a document whose entire image is composed of a photographic image or the like (hereinafter referred to as “multi-density target image”) are discriminated.

Then, in step S503, electronic filing processing is performed. Specifically, the input multi-valued image data is binarized to perform image area separation, and the text area is subjected to OCR processing according to the image area separation result. Also, the binarized image is
Filing registration is performed in the storage device 103 together with information that characterizes images for image retrieval. These processes are performed according to the type of image determined in step S502. In other words, the "single density object image" is simply binarized to separate the image area, the text area is subjected to OCR processing, and the obtained binary image and character code are registered in filing. On the other hand, the "multi-density target image" is obtained by first binarizing optimally for image area separation, removing the background, then separating the image area, and performing simple binarization on the character area. The registered binary image and character code are registered in filing. In addition, pseudo halftone binarization is performed on the image area by the error diffusion method or the dither method, and the obtained binary image is registered as filing.

FIG. 3 is a flow chart for explaining an example of the image discrimination processing executed by the image processing apparatus 102 of the present embodiment in detail.

In FIG. 3, in step S1, the image input device 1
A multi-valued image input from 01 and temporarily stored in the storage device 103 is input to the memory. Then, the histogram of the multi-valued image input in step S2 is calculated. Here, the frequency of the brightness value is calculated for all pixels of the image.

Next, in step S3, the histogram is divided into two with the average value of the obtained histogram as a boundary, and the divided histogram is further divided into two with the average value of each histogram to obtain four histograms. . This division is n
Repeated times, we get 2 ^ n (n to the power of 2) histograms in a narrow luminance interval. In this embodiment, the histogram is divided into eight histograms with n = 3.

Next, the skew statistic Sk of each histogram divided in step S4 is calculated. Here, the Sk value is a value that represents the bias of the frequency distribution, and is calculated using equation (1). SK = (Σ (Xi-AV) ^ 3) / D… (1) D = Σ (Xi-AV) ^ 2… (2) where Xi: Population element (pixel brightness) D: Whole population Variance of AV: population mean a ^ b: a to the power b

The calculation formula of the Sk value includes the cube of the difference between the population and its average value AV. From this, it can be seen that when the sign of the Sk value is positive, there is a bias of the population on the side larger than the average value AV, and as the Sk value increases, the average value AV also shifts toward the larger deviation. . That is, the sign of the Sk value indicates the bias direction of the luminance distribution, and the value indicates the degree of the bias. Note that the minimum brightness value (that is, the black side)
As you can see, the Sk value of each histogram with a small brightness
(1), Sk (2),…, Sk (2 ^ n).

Next, in step S5, which will be described in detail later,
The shape of the human gram is judged from the obtained eight Sk values,
Output the number of valleys (minimum) in the histogram. Subsequently, the type of image is determined from the number of valleys obtained in step S6.
Fig. 4 shows an example of a luminance histogram when a simple document is read by a scanner.When an image with such a histogram is input and it is judged that the number of valleys is one, "single density target image Is determined. Figure 5 is an example of a luminance histogram when a document or drawing with a density change in the background is read by a scanner.When an image with such a histogram is input and it is determined that there are one or more valleys, It is determined to be a “multi-density target image”.

FIG. 6 is a flow chart showing an example of processing for checking the number of valleys in the histogram in step S5.

First, in step S151, an initial value 0 is substituted into the luminance section number k and the number i of valleys. Then, in step S152, Sk
(k)> 0 is determined, and Sk (k + 1) <0 is determined in step S153. If both are true, the number i of valleys is incremented in step S154. That is, when the Sk value changes from positive to negative, it is determined that there is a valley, 1 is added to the number i of valleys in step S154, and otherwise the processing jumps to step S155.

Next, in step S155, it is determined whether or not all the sections of the luminance distribution have been checked by determining k + 1 <2 ^ n. If all the sections have been checked, the process is terminated. If not, the luminance section number k is incremented in step S156, and the process returns to step S152.

By the above procedure, the shape of the histogram can be judged by checking the number of valleys. Next, a method of discriminating an image type will be described by actually showing numerical values.

FIG. 7 is a diagram showing an example of a change in the brightness section when dividing the histogram shown in FIG. 4 and an Sk value in each section. The first interval is 0 to 255 because it is assumed to be an 8-bit multi-valued image. The average value AV in this section is 177, and the section is divided into 0 to 176 and 177 to 255. Further, the average value AV of the sections 0 to 176 is 91, and the section is divided into 0 to 90 and 91 to 176. Divide in the same way below,
Eight intervals (k = 1 to 8) are obtained.

Then, the Sk (1) value in the section k = 1 (0 to 42) is 7.0.
8, the k (2) value of the section k = 2 (43 to 90) is -7.00, and so on.
(3) = 0.67, Sk (4) = 0.94, Sk (5) =-4.46, Sk (6) =-1.04, Sk
(7) = 0.57 and Sk (8) = 6.12 are obtained. Among these Sk values, the negative value is changed to the positive value when viewed from the direction of the minimum luminance between Sk (2) and Sk (3) (-7.00 to 0.67) and S
There are two places between k (6) and Sk (7) (-1.04 to 0.57).
That is, it can be determined that there are two valleys in this histogram, and as a result, the image shown in FIG. 4 is determined to be a multi-density target image.

As described above, according to this embodiment,
For example, the frequency of the brightness of the input multi-valued image is obtained, and the bias of the frequency distribution of each histogram obtained by repeatedly dividing the histogram based on the average value thereof, that is, the Sk value is calculated,
By determining the shape of the histogram from the change in the Sk value and determining the type of the multi-valued image, it is possible to perform electronic filing in which image processing is performed according to the type of image. For example,
When filing a document containing a photographic image, it is possible to perform a binarization process (such as an error diffusion method) for images with multiple densities to file a pseudo-halftone image that retains the gradation of the photographic image. It will be possible. Therefore, even when filing a large number of images, the type of the image can be automatically determined, so that the user does not need to confirm the type of each image.

[0032]

[Modification] In the above-described embodiment, the input image is an 8-bit gray scale, but the present invention is not limited to this, and a color image can be input. That is, as long as it is multi-valued data, it does not matter whether the number of bits is large or small.

In the above-described embodiment, the example in which the input image is temporarily stored in the storage device 103 and then the image processing is executed has been described. However, the histogram may be directly obtained by prescanning the input image. . Further, sampling at the time of calculating the histogram may be all pixels or every few pixels, and is not particularly limited. Furthermore, the calculation of the average value and the statistic Sk, etc. does not have to be performed in 8 bits, and can be performed with a smaller number of bits for speeding up, memory deletion, etc.

Further, in the above-described embodiment, the number of divisions of the histogram is set to 2 ^ 3 (8 divisions), but this is empirically determined, and it is determined according to the bit depth of the input image. It is set appropriately and is not limited to this. Further, although an example in which the histogram is divided into histograms of small luminance sections by the average value has been described, the present invention is not limited to this. You may divide by pitch.

Further, in the above-mentioned embodiment, two kinds of images are used, but the present invention is not limited to these two kinds, and may be further classified into three kinds, four kinds and more. The types of images are not limited to the “single density target image” and the “multi-density target image”, and other types may be set.

The present invention may be applied to either a system composed of a plurality of devices or an apparatus composed of a single device.

Further, it goes without saying that the present invention can be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0038]

As described above, according to the present invention,
It is possible to provide an image processing apparatus and method for determining the type of an input image.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an electronic filing system including an image processing apparatus according to an embodiment of the invention.

FIG. 2 is a flowchart showing an example of a procedure of electronic filing performed by the image processing apparatus shown in FIG.

FIG. 3 is a flowchart illustrating in detail an example of image determination processing executed by the image processing apparatus shown in FIG.

FIG. 4 is an example of a luminance histogram when a simple document is read by a scanner,

FIG. 5 is an example of a luminance histogram when a document or drawing having a density change in the background is read by a scanner,

FIG. 6 is a flowchart showing an example of processing for checking the number of valleys in the histogram in step S5 shown in FIG.

7 is a diagram showing an example of a change in a brightness section when dividing the histogram shown in FIG. 4 and an Sk value in each section.

[Explanation of symbols]

 101 image input device 102 image processing device of this embodiment 103 storage device 104 image output device

Claims (15)

[Claims] 1. A calculation step of obtaining a luminance distribution from an input image, a division step of dividing the luminance distribution obtained in the calculation step into a plurality of luminance sections, and a bias of each of the luminance distribution divided in the dividing step. And a determining step of determining the type of the input image. 2. A first calculation step of obtaining an overall luminance distribution from an input image, a division step of dividing the entire luminance distribution obtained in the first calculation step into a plurality of luminance sections, and a division of the division step. The second calculation step for calculating the value representing the bias of each of the luminance distribution, the determination step for determining the shape of the overall luminance distribution based on the calculation result of the second calculation step, and the determination step And a determination step of determining the type of the input image from the shape of the overall luminance distribution. 3. The dividing step divides the entire luminance distribution into a plurality of luminance sections by repeating a step of dividing the luminance distribution into two with an average value of the luminance distribution as a boundary, a predetermined number of times. The image processing method according to claim 2. 4. The second calculation step calculates a value representing the bias of the brightness distribution from the difference between each sample value included in the brightness distribution and the average value of the brightness distribution. The image processing method described in item 2. 5. The image processing method according to claim 4, wherein the second calculation step calculates a value representing the bias of the brightness distribution by using an odd power of the difference. 6. The determination step, based on the number of times the value representing the bias of the divided luminance distribution changes from positive to negative from the minimum value side to the maximum value side of the luminance section,
4. The image processing method according to claim 2, wherein the shape of the entire luminance distribution is determined. 7. The method according to claim 6, wherein the determining step determines that the brightness distribution has a local minimum in the vicinity of a brightness section in which a value representing the bias of the brightness distribution changes from positive to negative. Image processing method. 8. The determining step determines that the input image is an image composed of a single density object when the number of changes obtained in the determining step is 1. 6. Image processing method described in. 9. The determination step determines that the input image is an image composed of a plurality of different density targets or continuous tone density targets when the number of changes obtained in the determination step is other than 1. 7. The image processing method according to claim 6, wherein. 10. The method further comprises a processing step of performing image processing on the input image based on the determination result of the determination step, and a storage step of storing the image processed in the processing step in a storage means. The image processing method according to any one of claims 1 to 9. 11. The processing step performs simple binarization on the input image if the discrimination result represents an image composed of a single density object, and if not, pseudo halftone processing is performed on the input image. 11. The method according to claim 10, wherein the binarization is performed.
Image processing method described in. 12. A calculating means for obtaining a brightness distribution of an input image, a dividing means for dividing the brightness distribution obtained by the calculating means into a plurality of brightness sections, and a bias of each of the brightness distributions divided by the dividing means. An image processing apparatus, comprising: a determining unit that determines the type of the input image based on the above. 13. A first arithmetic means for obtaining an overall luminance distribution from an input image, a dividing means for dividing the overall luminance distribution obtained by the first arithmetic means into a plurality of luminance intervals, and the dividing means. Second computing means for computing a value representing the bias of each of the divided luminance distributions, determining means for determining the shape of the overall luminance distribution based on the computation result of the second computing means, and the determining means An image processing apparatus, comprising: a determining unit that determines the type of the input image based on the determined shape of the overall luminance distribution. 14. Input means for inputting the input image, processing means for performing image processing on the input image based on the discrimination result of the discriminating means, and an image processed by the processing means is stored. 11. The storage device according to claim 11 or 12,
The image processing device described in. 15. The processing means performs simple binarization on the input image when the determination result represents an image composed of a single density object, and when not, pseudo halftone processing is performed on the input image. 15. The image processing device according to claim 14, wherein binarization is performed.