JP4001446B2 - Method, apparatus and computer-readable recording medium for specifying image background color - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a document image recognition method, and more particularly to a document image recognition method for specifying a background color of a color document image.
[0002]
[Prior art]
What is ultimately required as a result of document image recognition is a layout configuration for correctly recognizing characters (contents), tables and ruled lines, diagrams, photographs, and reading order. There is no need to reproduce exactly. Therefore, it has not been so much considered in the past to accurately specify the background color of a color document image.
[0003]
However, by removing the background color of the document, it may be possible to extract a necessary document element, and the background color can be one of important features for document search using image features. Therefore, a technique for specifying the background color of a color document image is important.
[0004]
As a conventional technique for specifying the background color of a color document image, Japanese Patent Application Laid-Open No. 06-348758 “Document Information Retrieval Apparatus and Method” discloses a document information inspection apparatus capable of retrieving a desired document using information representing the appearance of the document as a retrieval key. In this document, the background color of the document is detected and held as one of the search keys.
[0005]
[Problems to be solved by the invention]
In the above prior art, the background color of the document is simply defined as the color that appears most frequently in the document. Therefore, although simple and high-speed processing is possible, there is a possibility that an erroneous color is set as a background color when a large figure / photo is present in the document.
[0006]
As described above, when a large figure / photo is present in the document, there is a problem that the cluster size of another color becomes larger than the background color cluster size as a result of color clustering, and this is used as the background color.
[0007]
The present invention has been made in view of the above points, and an object thereof is to provide a method and apparatus for accurately specifying a background color even when a large figure / photograph exists in a document.
[0008]
Further, in consideration of the fact that the background color specification of a color document image is a preprocessing for obtaining the result of document image recognition, a method and apparatus for specifying a background color that can be processed simply and at high speed The purpose is to provide.
[0009]
[Means for Solving the Problems]
In order to solve the above problems, the present invention is configured as follows.
[0010]
The invention according to claim 1 is a method executed by an apparatus for specifying a background color of a document image input as a digital image, the clustering step for clustering colors in the document image, and the clustering step. The color of the first cluster having the largest number of pixels among the plurality of obtained clusters and the color of the second cluster having the second largest number of pixels among the plurality of clusters are determined. The determination step, and the difference between the number of pixels of the first cluster and the second cluster is smaller than a predetermined value, and the color of the second cluster is distributed over the entire document image. A background color specifying step in which the color of the two clusters is used as the background color of the document image.
[0011]
From the viewpoint that the background color is distributed over the entire document, when the second color is distributed over a wide range in the document, the background color is used as the background color. Even when there is a photo, the background color can be obtained with high accuracy. If the above condition is not satisfied, the maximum cluster may be used as the background color, and the background color can be obtained at high speed.
[0014]
The invention according to claim 2 is an apparatus suitable for carrying out the above method, and has the same effect as the above effect.
[0015]
The invention described in claim 3 is a recording medium on which a program suitable for executing the above method on a computer is recorded, and has the same effect as the above effect.
[0016]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram of a document image recognition apparatus according to an embodiment of the present invention. As shown in the figure, the document image recognition apparatus of the present invention includes a CPU (central processing unit) 100, a memory 101, a scanner 102, a communication device 103, a display device 104, a hard disk 105, a keyboard 106, and a CD-ROM drive 107. Have. The hard disk 105 may be connected as an external apparatus of the document image recognition apparatus or may be included as an internal apparatus. The CPU 100 controls the entire document image recognition apparatus. The memory 101 holds data and programs processed by the CPU 100. The communication device 103 performs control for connecting the document image recognition device to a network such as a LAN. The keyboard 106 is a device for inputting data. The CD-ROM drive 107 drives a CD-ROM or the like to read / write. The scanner 102 is connected as an external device of the document image recognition device and has a function of inputting a document as a color digital image. In addition to the CD-ROM drive 107, an MO drive, an FD drive, a DVD drive, or the like may be used.
[0017]
A program for executing processing to be described later in the embodiment of the present invention is stored in, for example, a CD-ROM and loaded onto the hard disk 105 via the CD-ROM drive 107. When the program is activated, a predetermined program is expanded in the memory 101 and processing is executed. The outline of the operation is as follows.
[0018]
First, the content of a paper document is input as a color digital image via the scanner 102 and stored in the memory 101 or the hard disk 105. Next, a region identification process including background color specification described in detail below is performed by the CPU 100, and for example, an OCR process is performed on the region identification result.
[0019]
Next, processing for specifying the background color will be described in detail with reference to the flowchart of FIG.
[0020]
In step S1, a paper document is input to the document image recognition apparatus as a color digital image by the scanner 102 or via a network. In this embodiment, it is assumed that the number of colors is 24 bit full color, but other colors and resolutions may be used. The present invention basically does not depend on the number of colors or resolution of an image. Smoothing is performed in step S2, and image compression is performed in step S3. The order of the smoothing and image compression processes may be reversed.
[0021]
Smoothing and image compression are processes performed as preprocessing for color clustering (step S4) for examining a rough color distribution in a document. A part of the color image is composed of halftone dots of different colors, and these halftone dot configurations are different depending on models such as printers and color copies. In order to make the characteristics of the color output system uniform, the entire image is smoothed. Therefore, smoothing may not be performed if not necessary. Further, since smoothing is generally a time-consuming process, it may not be performed even when speed is a problem. Image compression is processing performed to reduce the access cost to images, and the compression rate for image compression can be changed dynamically during operation. Similar to smoothing, image compression may not be performed.
[0022]
In step S4, color quantization is performed under an appropriate color coordinate system (for example, RGB), and color clustering is performed on the entire image in order to examine a rough color distribution. This color clustering is performed by a general technique in the image processing field with the final number of clusters set to about 2 to 10.
[0023]
In step S5, representative colors of the largest cluster (referred to as A) and the second highest cluster (referred to as B) are determined. There are the following two methods for determining the representative color of the cluster.
[0024]
(1) The average value of each RGB plane is obtained and used as a representative color.
[0025]
(2) Taking a histogram of luminance, a color corresponding to the median value is set as a representative color.
[0026]
In general, the method of obtaining the average value of (1) is advantageous in terms of speed, and the method of obtaining the median value of (2) is advantageous in terms of accuracy.
[0027]
Next, a background color is specified (steps S6 to S10). In general, since the number of background pixels is sufficiently larger than the number of other color pixels, the representative color of the cluster having the largest number of belonging pixels is set as the background color of the document.
[0028]
However, as shown in FIG. 3, when a document having a figure / photo larger than the background exists, the number of pixels belonging to another color cluster is larger than the number of pixels belonging to the color cluster corresponding to the background color. Sometimes. In order to deal with such a case with high accuracy, in the present invention, when the size of the maximum cluster C1 is S1 and the size of the second-ranked cluster C2 is S2, a threshold value having a difference between S1 and S2 is used. If C1 and C2 satisfy a specific condition, the representative color of the second highest cluster C2 is set as the background color of the document.
[0029]
That is, if the number of pixels belonging to a certain cluster X is called the size of the cluster X and is described as N (X), if N (A) is sufficiently larger than N (B) in step S6, The representative color of cluster A is determined as the background color of the document (step S7), and the process ends. Here, “sufficiently large” specifically means, for example, a case where there is an opening of 1.5 times or more. That is, N (A)> N (B) × 1.5.
[0030]
In step S6, when N (A) is not sufficiently larger than N (B), the distribution degree of the pixels belonging to each cluster in the document is examined. In step 8, if the pixels belonging to cluster B exist in the document as large connected components, that is, if the representative color of cluster B is distributed over a wide range of the document, the representative color of the cluster is set as the background color of the document. (Dotted line from step S8 to S10). FIG. 4 is a diagram showing the color distribution in the document of FIG. 3, and is an example when the color B is distributed over a wide range. On the other hand, if the representative color of cluster B is not widely distributed in the document, that is, if the pixels belonging to cluster B exist as a large number of unconnected small clusters, the representative color of cluster B is the background of the document. Since it is determined that it cannot be a color, the representative color of cluster A is determined as the background color of the document (step S7). Here, “the representative color of the cluster B is distributed over a wide range of the document” specifically means that, for example, when a histogram of the representative color of the cluster B is taken for each line in the horizontal direction, the interval of the frequency 0 Can be assumed to be 50% or less.
[0031]
In step S8, if the background color cannot be determined even if the representative color of cluster B is distributed over a wide range of the document, it is checked in step 9 whether cluster B is whitish than cluster A. The representative color of the more whitish cluster is determined as the background color of the document (step S7 or step S10). This is because the background color is considered whitish than the colors other than the background. Here, “color B is whitish than color A” specifically means that the luminance of color B is larger than the luminance of color A. In the RGB system, white is represented as (255, 255, 255), and black is represented as (0, 0, 0). Therefore, for example, a larger Green value can be defined as more whitish.
[0032]
In the above description, the cluster B is the second-ranked cluster. However, if the cluster is smaller than the cluster A, the cluster B need not be the second-ranked cluster. B is selected as the background color as long as the conditions of NO in step S6, YES in S8, and YES in S9 are satisfied.
[0033]
In step S6, if N (A) is not sufficiently larger than N (B), the subsequent processing is not performed, and even if a warning is given to the user that it is difficult to automatically specify the background color at this time. Good. For example, a warning such as “Failed to specify background color automatically” is issued.
[0034]
The present invention is not limited to the above-described embodiments, and various modifications and applications can be made within the scope of the claims.
[0035]
【The invention's effect】
As described above, according to the present invention, the background color of a color document image can be obtained at high speed and with high accuracy. In addition, the background color can be accurately specified even when a large figure or photograph exists in the document. Therefore, the image recognition process using the background color can be performed efficiently.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a document image recognition apparatus according to an embodiment of the present invention.
FIG. 2 is a flowchart of background color specifying processing according to an embodiment of the present invention.
FIG. 3 is a diagram illustrating an example of a document including a photograph larger than the background.
FIG. 4 is a diagram showing color distribution in the document of FIG.
100 CPU
101 Memory 102 Scanner 103 Communication Device 104 Display Device 105 Hard Disk 106 Keyboard 107 CD-ROM Drive

Claims (3)

A method performed by an apparatus for specifying a background color of a document image input as a digital image , comprising:
A clustering step of clustering colors in the document image;
The color of the first cluster having the largest number of pixels among the plurality of clusters obtained by the clustering step, and the second cluster having the second largest number of pixels among the plurality of clusters. A decision step to determine the color;
When the difference in the number of pixels between the first cluster and the second cluster is smaller than a predetermined value and the color of the second cluster is distributed over the entire document image, the second cluster And a background color specifying step of setting a color as a background color of the document image. An apparatus for specifying a background color of a document image input as a digital image ,
Clustering means for clustering colors in the document image;
The color of the first cluster having the largest number of pixels among the plurality of clusters obtained by the clustering means, and the second cluster having the second largest number of pixels among the plurality of clusters. A determining means for determining a color;
When the difference in the number of pixels between the first cluster and the second cluster is smaller than a predetermined value and the color of the second cluster is distributed over the entire document image, the second cluster And a background color specifying means for setting a color as a background color of the document image. A computer-readable recording medium recording a program for executing processing for specifying a background color of a document image input as a digital image to a computer,
A clustering procedure for clustering colors in the document image;
The color of the first cluster having the largest number of pixels among the plurality of clusters obtained by the clustering procedure, and the second cluster having the second largest number of pixels among the plurality of clusters. A decision procedure to determine the color;
When the difference in the number of pixels between the first cluster and the second cluster is smaller than a predetermined value and the color of the second cluster is distributed over the entire document image, the second cluster A computer-readable recording medium storing a program for causing a computer to execute a background color specifying procedure for setting a color as a background color of the document image.

Images