JP3756660B2 - Image recognition method, apparatus and recording medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image recognition method , apparatus, and recording medium for accurately extracting a predetermined mark from an input color image.
[0002]
[Prior art]
Various methods have been proposed as image recognition methods for color images. For example, the present applicant needs to perform image recognition when performing color image recognition processing on a color image, in particular, a pattern in which an image to be extracted is drawn according to a predetermined rule. A color image recognition method has been proposed that recognizes an object at high speed by compressing the amount of data without affecting the rotation while ensuring a large amount of information (see Japanese Patent Laid-Open No. 8-263661).
[0003]
[Problems to be solved by the invention]
In the image recognition method described above, when a predetermined image signal is extracted from the input image signal to generate a binarized image, the predetermined image is generated from the input color image signal (R, G, B). Signal components are extracted to generate a binary image.
[0004]
However, in this method, a binary image is generated only by the level of the image signal for each pixel without considering the difference from the image density of the background of the image to be recognized (for example, printed paper). There is a problem that even if there is a density difference from the background, the mark cannot be extracted when the image signal level does not reach a predetermined threshold value. In addition, since a binarization table is not used, detailed threshold setting cannot be performed, or when detailed threshold setting is performed, many conditions are required, and there is a problem that processing speed decreases.
[0005]
An object of the present invention is an image recognition method for accurately extracting a predetermined mark by considering not only the level of an image signal for each pixel but also a density difference with neighboring pixels when generating a binary image. An apparatus and a recording medium are provided.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, according to the first aspect of the present invention, a predetermined image signal component is extracted from an input color image signal to generate a binarized image, and the binarized image is used to generate the binarized image. Extracting a predetermined mark, extracting a predetermined pattern composed of the plurality of extracted marks, and comparing the extracted pattern with a dictionary in which a pattern to be recognized is registered An image recognition method for recognizing a pattern, wherein when a binary image is generated from the input color image signal, a two-dimensional or higher binarization table referenced by the density of each pixel constituting the image signal is used. It is characterized by that.
[0007]
The invention according to claim 2 is characterized in that the two-dimensional or higher binarization table is referred to by the density of each pixel and the density of pixels in the vicinity thereof.
[0008]
The invention according to claim 3 is characterized in that the density of the neighboring pixels is calculated from data of one row in the main scanning direction.
[0009]
According to the fourth aspect of the present invention, means for extracting a predetermined image signal component from the input color image signal to generate a binarized image, and extracting a predetermined mark from the binarized image The extracted pattern by comparing the means, the means for extracting a predetermined pattern composed of the plurality of extracted marks, and the dictionary in which the extracted pattern is registered with the pattern to be recognized An image recognition apparatus comprising a unit for recognizing image data, wherein when generating a binary image from the input color image signal, a two-dimensional or higher binarization table referred to by the density of each pixel constituting the image signal It is characterized by using .
[0010]
According to a fifth aspect of the present invention, there is provided a computer-readable recording medium on which a program for causing a computer to realize the image recognition method according to any one of the first to third aspects is recorded.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings.
<Example 1>
FIG. 1 shows the configuration of an embodiment of the present invention. In the figure, 1 is an image input unit that optically reads a document to be recognized and the like, 2 is an image binarization unit that creates a binarized image from a color image input by the image input unit, and 3 is an image binarization unit. , 4 is a basic mark extraction unit for extracting a basic mark from a binarized image, 5 is a basic pattern extraction unit for extracting a basic pattern composed of a plurality of basic marks, and 6 is an extracted basic A basic pattern recognizing unit that collates a pattern with a dictionary, 7 is a determining unit that determines whether an input document is a recognition target document based on the recognized basic pattern, and 8 is an output unit that outputs a determination result. .
[0012]
FIG. 2 shows an example of a document to be recognized and an example of its basic pattern. As shown in FIG. 2, the present embodiment will be described by taking as an example a case where a repetitive pattern printed on the entire surface of the recognition target document 21 such as a ticket is detected and recognized. The basic pattern 23 constituting the repetitive pattern will be described by taking as an example a pattern having a basic mark (black circle) 22 at each vertex of an equilateral triangle indicated by a broken line in the drawing.
[0013]
First, the image binarization unit 2 excludes a character portion that is not a repetitive pattern to be detected in the example of the document in FIG. 2 (for example, extracts pixels having a value equal to or greater than a predetermined threshold), and repeats the pattern. Only binary images are generated.
[0014]
Next, the basic mark extraction unit 4 extracts a basic mark (in this example, a black circle) by using the method described in the image recognition method of the above-mentioned application, obtains its center coordinate, and the basic pattern extraction unit 5 Finds the center of the basic pattern (regular triangle in this example). Then, the basic pattern recognition unit 6 uses the distance from the center of the basic pattern to the basic mark to collate the distance of the basic pattern candidate with the distance in the dictionary, so that the basic pattern candidate is the pattern to be extracted. It is determined whether or not there is.
[0015]
The determination unit 7 determines whether or not the document is a recognition target document by using a condition that the number of basic patterns extracted as described above is a predetermined number or more in the target document.
[0016]
In this embodiment, in the above-described processing, the image binarization unit 2 binarizes the input color image using a binarization table as shown in FIG. In this example, description will be made using a two-dimensional binarization table, but a multi-dimensional table such as three-dimensional (RGB or the like) or four-dimensional (YMCK or the like) may be used.
[0017]
For example, when binarization is performed using values of R (0 to 255) and G (0 to 255) from an input image signal (each color is represented by 8 bits), FIG. As shown in the figure, a matrix is created with the values of R and G, and 1 or 0 is set for each element of the matrix.
[0018]
That is, for each pixel of image data obtained by reading a document to be recognized, the pixel value (R, G, B) of a certain pixel i is (R, G, B) = (246, 66, 19). Referring to the binarization table of FIG. 3, since the value of the element of the matrix where R = 246 and G = 66 is 1, binarization is performed with the pixel i as a black pixel. Such processing is performed for all the pixels constituting the image of the document to be recognized.
[0019]
<Example 2>
In the first embodiment, the binarization table is referred to by using the R and G values of the pixel values of the input image. However, in this embodiment, nearby pixel values are also used. FIG. 4 shows the configuration of the second embodiment of the present invention. The image binarization unit 32 includes a filter processing unit 33 and a binarization table 34. Other components are the same as those in the first embodiment.
[0020]
The filter processing unit 33 processes as follows. In other words, the G density of the pixel of interest to be binarized is now G (X0, Y0). At this time, as shown in FIG. 5, for example, the pixel values of the reference pixels G (X0-3, Y0) and G (X0 + 3, Y0) separated by 3 pixels on the left and right in the main scanning direction are also referred to. 0, 0, 2, 0, 0, -1 "is performed (-1 * G (X0-3, Y0) + 2 * G (X0, Y0) -1 * G (X0 + 3, Y0)) As a result, a differential value of the concentration of G can be obtained. The differential value of B concentration is obtained in the same manner.
[0021]
By using the G and B differential values and referring to the binarization table 34 as described in the first embodiment, “the differential value is equal to or greater than a predetermined value and the density value is within a predetermined range. ”Can be obtained.
[0022]
<Example 3>
An example of calculating the density difference from the above-described neighboring pixels in the sub-scanning direction will be described with reference to FIG. That is, for example, in order to refer to the pixel values of the previous row G (X0, Y0-1) and the subsequent row G (X0, Y0 + 1) with the G density of the target pixel being G (X0, Y0). Requires a memory for storing the data of the (Y0-1) th and (Y0 + 1) th lines in addition to the image data of the current line Y0. That is, the required amount of memory increases when the density difference with neighboring pixels is obtained in the sub-scanning direction.
[0023]
Therefore, in this embodiment, by calculating the density difference with the neighboring pixels from the data of one row in the main scanning direction, the input signal data for one row that is currently processed may be stored in the memory. That is, the line memory is not required, and the amount of memory can be reduced.
[0024]
<Example 4>
FIG. 7 shows an embodiment in which the present invention is realized by software. When the present invention is implemented by software, as shown in FIG. 7, a computer system including a CPU, a memory, a display device, a hard disk, a keyboard, a CD-ROM drive, a scanner, etc. is prepared, and a computer such as a CD-ROM is prepared. A readable recording medium stores a program for realizing the image recognition function of the present invention, a pattern dictionary, and the like. A document image to be processed is stored in a hard disk or the like. When the program is started, the original image data is read, image recognition processing is executed, and the recognition result is output to a display or the like.
[0025]
【The invention's effect】
As described above, according to the first, fourth , and fifth aspects of the invention, since a binarization table is used when generating a binary image, detailed threshold setting can be performed, and high speed can be set. A binary image can be generated.
[0026]
According to the second aspect of the present invention, since a density difference with neighboring pixels is used when generating a binary image, if there is a density difference between the background and the mark to be extracted, the binary image is accurately extracted. can do.
[0027]
According to the third aspect of the present invention, since the density difference with the neighboring pixels is calculated from one line of data in the main scanning direction, the input image data for one line currently being processed is stored in the memory. What is necessary is to reduce the amount of memory used.
[Brief description of the drawings]
FIG. 1 shows a configuration of Embodiment 1 of the present invention.
FIG. 2 shows an example of a document to be recognized and an example of its basic pattern.
FIG. 3 shows an example of a binarization table.
FIG. 4 shows a configuration of Embodiment 2 of the present invention.
FIG. 5 is a diagram illustrating a process according to the second embodiment.
FIG. 6 shows an example of calculating a density difference with neighboring pixels in the sub-scanning direction.
FIG. 7 shows a configuration example when the present invention is realized by software.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 Image input part 2 Image binarization part 3 Binarization table 4 Basic mark extraction part 5 Basic pattern extraction part 6 Basic pattern recognition part 7 Determination part 8 Output part

Claims (5)

A predetermined image signal component is extracted from the input color image signal to generate a binarized image, a predetermined mark is extracted from the binarized image, and a plurality of extracted marks are included. An image recognition method for recognizing the extracted pattern by extracting the predetermined pattern and comparing the extracted pattern with a dictionary in which a pattern to be recognized is registered, wherein the input color image An image recognition method characterized by using a two-dimensional or higher binarization table referenced by the density of each pixel constituting the image signal when generating a binary image from the signal. 2. The image recognition method according to claim 1, wherein the two-dimensional or higher binarization table is referred to based on the density of each pixel and the density of pixels in the vicinity thereof.   The image recognition method according to claim 1, wherein the density of the neighboring pixels is calculated from data of one row in the main scanning direction. Means for extracting a predetermined image signal component from the input color image signal to generate a binarized image; means for extracting a predetermined mark from the binarized image; An image provided with means for extracting a predetermined pattern composed of the mark, and means for recognizing the extracted pattern by collating the extracted pattern with a dictionary in which a pattern to be recognized is registered An image recognition apparatus, wherein when generating a binary image from the input color image signal, an image using a two-dimensional or higher binarization table referenced by the density of each pixel constituting the image signal is used. Recognition device. The computer-readable recording medium which recorded the program for making a computer implement | achieve the image recognition method of any one of Claim 1 thru | or 3 .

Images