JP5204035B2 - Image processing apparatus, processing method, and processing program - Google Patents

Description

  The present invention relates to an image processing apparatus, a processing method, and a processing program for automatically dividing and storing a small image from a medical image in which a plurality of small images are recorded on a single film, and particularly to each small image. The present invention also relates to an image processing apparatus, a processing method, and a processing program that can automatically divide and store character information and the like recorded together with each small image.

  Conventionally, a medical image often has a plurality of small images recorded on a single film. When electronically storing the images as image data, the medical images are divided and stored for each small image, not for each film. It is usually done.

  Since the manual division of each small image requires a lot of labor, an apparatus for automatically dividing the small image into small images has been desired. In view of this, an image processing apparatus that automatically divides and stores a plurality of small images from such a single film has been proposed (see Patent Document 1). This image processing apparatus scans a film, captures an image, detects an arrangement of small images based on dispersion of pixel values constituting the image, and automatically divides a plurality of small images in a short time. It is something that can be done.

JP 2000-342539 A

  However, in the above image processing apparatus, when character information or the like accompanying the small image is recorded around a plurality of small images to be divided, the background portion between adjacent small images is almost transparent. In such a case, there is a problem that the character information or the like cannot be accurately detected and cannot be accurately divided in such a manner that the character information or the like is included in the small image.

  The present invention has been made in view of such circumstances, and when character information or the like is recorded around the small image, the small image including the character information or the like is not affected by the background portion. It is an object of the present invention to provide an image processing apparatus, a processing method, and a processing program that can be automatically and accurately divided.

The above problem can be solved by the following inventions.
That is, the image processing apparatus of the present invention includes an image capturing unit that captures an image, background detection means that detects a background portion other than an image in the captured image area, and image information included in the captured image. An image processing apparatus comprising: a plurality of small images, and an image dividing unit that divides character information associated with the small images for each small image, wherein the background detecting unit A background portion is detected from the brightness frequency distribution of the captured image, and the image dividing means uses the brightness frequency distribution of the non-background portion from the image area other than the background portion detected by the background detection means. A band-like region formed by dividing the captured image into the vertical direction and the horizontal direction including the small image and the character information using a frequency distribution of brightness other than the background portion. An area that is detected as an in and horizontal scan line, and the vertical scan line and the horizontal scan line are overlapped and overlapped as an entire image area seed for extracting all image areas including the plurality of small images. An area having a uniform area in which each small image accompanied by the character information is included in the entire image area by setting a boundary surrounding the outermost periphery of the entire image area seed in a rectangle and setting it as an entire image area A provisional dividing line that equally divides the entire image area vertically and horizontally from the arrangement of the entire image area and the small image to be arranged in the vicinity of the center. A region seed is divided to divide each small image associated with the character information, and the small image region is detected by synthesizing the small image and the associated character information.

Further, the background detection means obtains a brightness frequency distribution of pixels constituting the captured image and an average value of the brightness frequencies, and the brightness having a frequency exceeding the average value is the brightness of the background portion. It is detected as background brightness.
Thereby, the part excluding the background lightness part can be detected as a part having image information.

  Further, the image dividing means obtains a frequency distribution of brightness other than the background brightness among the brightness of the pixels constituting the captured image, and presets in the image area of the captured image based on the frequency distribution The arrangement of the small images is obtained from the frequency of brightness exceeding a set value.

  The background detection means obtains periodicity in all image region seeds and excludes a single title portion having no periodicity in all image region seeds.

The image dividing means determines that the interval between adjacent small image regions is narrow when the temporary dividing line passes through the entire image region seed, and determines all regions of all image region seeds of the detected small image. A small image area is divided into equal intervals according to the number.
As a result, even if the image has a small image interval, the image can be reliably divided and the small image region can be detected.

  An image processing method of the present invention includes an image capturing unit that captures an image, background detection means that detects a background portion other than an image in a region of the captured image, and a plurality of pieces of image information included in the captured image. A method of processing an image using an image processing apparatus comprising: an image processing unit comprising: an image processing unit configured to divide the small image and character information associated with the small image for each small image; A background portion is detected from a brightness frequency distribution of the captured image by a background detection means, and a lightness frequency distribution other than the background portion is detected from the image region other than the background portion detected by the background detection means by the image dividing means. The small image is detected using a band, and the captured image is divided into a vertical direction and a horizontal direction including the small image and the character information using a frequency distribution of brightness other than the background portion. For detecting a region as a vertical scan line and a horizontal scan line, and extracting the entire image region including the plurality of small images by overlapping the vertical scan line and the horizontal scan line. Detected as a whole image area seed, set a boundary that encloses the outermost periphery of the whole image area seed in a rectangle, and set the inside as a whole image area, and each small image accompanied by the character information in the whole image area A temporary dividing line for equally dividing the entire image area vertically and horizontally is set from the entire image area and the arrangement of the small images so as to be arranged in the vicinity of the center with a uniform area. The whole image region seed is divided by lines, each small image accompanied by the character information is divided, and the small image region is detected by synthesizing the small image and the accompanying character information.

  An image processing program according to the present invention is an image processing program for dividing a plurality of small images having image information from a captured image and character information associated with the small images for each small image, and an image capturing unit for capturing an image And background detecting means for detecting a background portion other than the image in the area of the captured image, a plurality of small images having image information included in the captured image, and character information associated with the small image as a small image An image processing unit comprising: an image dividing unit that divides each image; and a step of detecting a background portion from a frequency distribution of brightness of the image captured by the background detecting unit, and the image division Detecting the small image from the image area other than the background portion detected by the background detection means using a frequency distribution of brightness other than the background portion; A band-like region formed by dividing the image captured by the image dividing unit into the vertical direction and the horizontal direction including the small image and the character information using a frequency distribution of brightness other than the background portion is formed as a vertical scan line. Detecting a horizontal scan line and a region where the vertical scan line and the horizontal scan line are overlapped and overlapped by the image dividing unit to extract an entire image region including the plurality of small images. Detecting as a whole image area seed, setting a boundary that surrounds the outermost periphery of the whole image area seed in a rectangle by the image dividing means, and setting the inside as a whole image area, and by the image dividing means In order to arrange each small image with the character information in the entire image area to have an area of an equal area and to be arranged near the center, Setting a temporary dividing line that equally divides the entire image area vertically and horizontally from the area and the arrangement of the small images; and dividing the entire image area seed by the temporary dividing line by the image dividing unit; Dividing each small image accompanied by character information, and combining the small image and the accompanying character information to detect a small image region.

  As described above, according to the image processing apparatus, the processing method, and the processing program of the present invention, when character information or the like is recorded around a small image, the character information or the like is not affected by the background portion. Including small images can be automatically and accurately divided.

1 is a system configuration diagram of an image processing apparatus according to an embodiment of the present invention. It is a figure which shows the image area of the division | segmentation object in the image processing apparatus of this invention. It is a figure which shows the image area of the division | segmentation object in the image processing apparatus of this invention. It is a figure which shows the image area of the division | segmentation object in the image processing apparatus of this invention. It is a flow which shows the flow of the image division process in the image processing apparatus of this invention. It is a flow which shows the flow of the high precision division | segmentation in the image processing apparatus of this invention. It is a flow which shows the flow of the small image extraction in the image processing apparatus of this invention. It is a figure which shows the frequency distribution of the brightness of the pixel which comprises the image in the image processing apparatus of this invention. It is detection explanatory drawing of the division | segmentation image core part in the image processing apparatus of this invention. It is explanatory drawing of the division | segmentation image core part expansion in the image processing apparatus of this invention. FIG. 6 is an explanatory diagram of enlargement rectangle combination in the image processing apparatus of the present invention. It is a figure which shows all the image area | region seed detection in the image processing apparatus of this invention. It is a figure which shows all the image area | region seed detection in the image processing apparatus of this invention. It is a figure which shows all the image area detections in the image processing apparatus of this invention. It is a figure which shows all the image area detections in the image processing apparatus of this invention. It is a figure which shows the provisional dividing line production | generation in the image processing apparatus of this invention. It is a figure which shows the synthetic | combination division | segmentation rectangle production | generation in the image processing apparatus of this invention. It is a figure which shows the synthetic | combination division | segmentation rectangle production | generation in the image processing apparatus of this invention. It is a figure which shows the small image area | region divided | segmented by the high precision division mode in the image processing apparatus of this invention. It is a flow which shows the flow of the medium precision division | segmentation in the image processing apparatus of this invention. It is a flow which shows the flow of the background lightness detection in the medium precision division | segmentation of the image processing apparatus of this invention.

  Hereinafter, embodiments for carrying out an image processing apparatus, a processing method, and a processing program of the present invention will be described in detail with reference to the drawings. 1 to 21 are diagrams illustrating embodiments of the present invention. In these drawings, the same reference numerals denote the same components, and the basic configuration and operation are the same. To do.

<Configuration>
FIG. 1 is a system configuration diagram of an image processing apparatus according to an embodiment of the present invention. An image processing apparatus 1 according to the present invention includes an image capturing unit 2 that captures an image by scanning a film and the like, and an image processing unit 3 that processes an image captured by the image capturing unit 2.

  The image capturing unit 2 has functions such as a film scanner and an image scanner, and can capture an image from an external storage device in addition to capturing an image from a film.

  The image processing unit 3 includes a background detection unit 4 and an image division unit 5, and the background detection unit 4 is a unit that detects a background part that is a frame other than the image part in the captured image region. The background portion is a region that divides a plurality of small images arranged in the image region in a lattice shape, and is an opaque or almost transparent portion having a spacing of several mm to several tens of mm.

  The image dividing means 5 can detect a plurality of small images arranged in the image area using the background portion detected by the background detecting means 4, cut out each small image, and save each as a separate file. It can be done.

  2 to 4 show examples of images to be divided by the image processing apparatus 1 of the present invention in the high-precision division mode. FIG. 2 shows an image with a narrow background portion between small images. Indicates an image with a title area at the top, and FIG. 4 shows an image with a wide background between small images. The image processing apparatus 1 of the present invention can reliably decompose a small image region for images having different patterns.

<Operation>
The operation of the image processing apparatus 1 according to the embodiment of the present invention will be described with reference to FIGS.

FIG. 5 is an operation flow diagram of the image processing apparatus 1 according to the present invention. In this flow, the number of minute information such as character information in the image captured by the image capturing unit 2 is preset by the background detection unit 4. Step 1 for determining whether or not there is more than the threshold value, Step 2 for dividing the image information in the high-precision division mode when the number of minute information is larger than the threshold value, and division in the high-precision division mode are rejected. Step 3 in which the image dividing means 5 determines whether or not it has been rejected, and Step 4 in which the image dividing means 5 divides in the medium precision division mode when the number of minute information is smaller than the threshold value in Step 1 above. The image dividing means 5 determines whether or not the dividing line penetrates the small image area as a result of the division, and the image dividing means 5 performs reject setting when the dividing line penetrates the small image. Or Consists Step 6 for the processing ends, and if not rejected in step 3, the process ends, respectively in the case of dividing line in step 5 is not penetrated the small image.
This flow shows a flow for selecting whether the image captured by the image processing unit 3 is to be divided in the high-precision division mode or the medium-precision division mode.

<High-precision division mode>
FIG. 6 is an operation flow diagram in which the background detection unit 4 and the image dividing unit 5 divide in the high-precision division mode in step 2 described above. This flow is based on the image captured by the image capturing unit 2 by scanning the film. Step 21 for extracting a small image, Step 22 for detecting an entire image region seed (described later), Step 23 for detecting an entire image region from the entire image region seed, and tentative dividing line determination from the entire image region Step 24, Step 25 for determining whether or not the temporary dividing line penetrates the entire image area seed, Step 26 for simply dividing the entire image area when penetrating, and a case where the temporary dividing line does not penetrate the entire image area seed In step 27 for generating a composite divided rectangle, in step 28 for determining whether the small image area is smaller than the entire image area beyond the threshold, In step 29 for determining whether or not to penetrate the image area, and in step 28, if the small image area is smaller than the entire image area by a threshold, and if the dividing line penetrates the small image area in step 29, reject setting is performed. And the process is terminated at step 30. If the dividing line does not penetrate the small image area at step 29, the process is terminated.
This flow shows a flow in which the image dividing unit 5 divides the image captured by the image capturing unit 2 in the high-precision division mode.

The small image extraction in step 21 is the same as the flow up to the detection of the divided image core part in the medium precision division mode in step 4 described later, and FIG. 7 and FIG. 11 will be described in detail below.
In step 31, the background detection unit 4 obtains the frequency distribution of the brightness of the pixels constituting the image captured by the image capturing unit 2 as shown in FIG. Next, in step 32, the background detection means 4 obtains an average value 31 of the brightness frequency, and sets the brightness having a frequency exceeding the average value 31 as the brightness of the background portion (hereinafter referred to as background brightness). Reference numeral 32 indicates a range of background brightness. If the background portion cannot be detected from the background brightness, the process ends.

  In step 34, the image dividing means 5 obtains a frequency distribution of brightness other than the background brightness 32 among the brightness of pixels constituting the captured image, and in the image area of the captured image based on the frequency distribution. The arrangement of the divided image core unit 43 is detected as follows.

  FIG. 9 is an explanatory diagram of detection of the divided image core unit 43, and the horizontal axis of the image area is described as the X axis and the vertical axis is described as the Y axis. The image dividing means 5 obtains the frequency of lightness other than the background lightness 32 for each fine strip-shaped region parallel to the Y axis of the captured image, and represents this as a histogram 41.

  Similarly, the image dividing means 5 obtains the frequency of lightness other than the background lightness 32 for each fine strip-shaped region parallel to the X axis, and represents this as a histogram 42.

In the histograms 41 and 42 obtained in this way, the image dividing means 5 obtains a strip-like region having a lightness frequency greater than a preset setting value 44, and divides the intersecting portion of the strip-like regions. The image core unit 43 is detected. The set value 44 may be such that a small image is detected, and minute information such as character information around the small image is not detected.
The number and arrangement of small images 17 (described later) are detected from the divided image core unit 43 thus obtained.

Next, in order to detect minute information such as character information recorded around the small image 17, all image region seeds in step 22, that is, all image regions including a plurality of small images in the captured image are displayed. The detection of the seed in the previous stage for detection will be described in detail below with reference to FIGS.
FIG. 12 shows an image captured by the image capturing unit 2 from a film. The background detection unit 4 detects a set value (small information is also detected in advance) as a portion having a lightness other than the background lightness as shown in FIG. Therefore, a strip-like region having a brightness frequency greater than the setting value 44 described above is obtained in the vertical and horizontal directions, and these are detected as the vertical scan line 11 and the horizontal scan line 12. Represents a state. In this state, the vertical scan line 11 and the horizontal scan line 12 are detected as a two-layer structure. The upper title portion 13 where the film title is recorded is also detected as one of the horizontal scan lines.

  FIG. 13 shows a state in which the background detection unit 4 combines the vertical scan line 11 and the horizontal scan line 12 detected in FIG. 12 on the same plane, and detects the overlapping area as the entire image area seed 14. Represents. Further, the background detecting means 4 extracts a group of regions having a large area in the overlapping region as a total of twelve small images 17 arranged in three rows in the horizontal direction and four rows in the vertical direction.

The entire image area detection in step 23 will be described in detail below with reference to FIGS.
FIG. 14 shows only one of the horizontal and vertical scan lines 11 and 12 that does not intersect in order to exclude the area in which no image is recorded from the vertical and horizontal scan line portions detected by the background detection means 4 in FIG. The horizontal scan line including the upper title portion 13 is also detected as having a different pattern from the other horizontal scan lines 12 having periodicity of image and character information. This represents a state immediately before detection of all image areas, in which all image area seeds 14 are extracted.

  FIG. 15 illustrates a state in which the background detection unit 4 sets a boundary that surrounds the outermost periphery of the entire image region seed 14 extracted in FIG. 14 in a rectangle and detects the inner side as the entire image region 16. Yes.

The provisional dividing line determination in step 24 will be described in detail below with reference to FIG.
In FIG. 16, the image dividing means 5 converts the small images 17 arranged in 3 rows horizontally and 4 rows vertically into the entire image area 16 detected by the background detecting means 4 up to FIG. It shows that it is divided evenly. Specifically, two vertical temporary dividing lines 18 that divide the horizontal width of the entire image area 16 into three at equal intervals, and three horizontal temporary divisions that divide the vertical width of the entire image area 16 into four at equal intervals. The line 19 is set.

The composite divided rectangle generation in step 27 will be described in detail below with reference to FIGS.
FIG. 17 shows a state in which the image dividing unit 5 generates a composite divided rectangle by superimposing the temporary dividing lines 18 and 19 in FIG. 16 and the entire image region seed 14 in FIG.

  FIG. 18 shows the small image 17 in which the image dividing means 5 is located in the area divided by the temporary dividing lines 18 and 19 in FIG. The small image 17 and the surrounding minute information area are combined into one image and detected as 12 small image areas 20.

  An image example of the small image area 20 detected in this way is shown in FIG. 19, in which the small image 17 is arranged at the center, and minute character information 21 accompanying the small image 17 is recorded around the small image 17. It is what. In order to cut out the character information 21 together with the small image 17 accurately, it is necessary that the character information 21 is positioned diagonally across the small image 17, and the arrangement of the character information 21 is one film. Therefore, it is necessary for the small images 17 to be somewhat similar to each other. This is because the image and character information must have periodicity in order for the background detection unit 4 to exclude the title part 13.

<Medium precision division mode>
FIG. 20 is an operation flow diagram in which the image dividing unit 5 divides in the medium-precision division mode in step 4 described above. This flow shows the background lightness in the image captured by the image capturing unit 2 by the background detection unit 4. A step 31 for detecting, and a step 32 for determining whether or not the background exists, that is, whether or not the background lightness has been detected. If it is displayed on the screen or the like and the process is terminated, and it is detected, the next step 33 is executed.

  In the flow following step 33, the background detection unit 4 determines whether or not the background brightness range is larger than a preset setting range. A step 34 for detecting a portion, a step 35 for retrying the background detection when it is larger than the set range, and a step 36 for determining whether or not the background lightness has been determined by the background detection means 4. If the background brightness cannot be determined in step 36, an error has occurred and the fact is displayed on the screen and the process is terminated. If the background brightness can be determined, step 34 is executed.

In the flow following step 34, the image dividing means 5 detects a rectangle in which the divided image core portion is enlarged, and the image dividing means 5 counts the number of enlarged rectangles 71 and combined rectangles 81 described later. Step 38 for determining whether or not the number is 1, and if the number is 1, the fact is displayed on the screen or the like to end the processing, and if the number is not 1, the image is divided. The means 5 comprises a step 39 of generating and checking a dividing line for dividing each small image in the image area. At this time, if the generated dividing line intersects the final rectangle, the image dividing unit 5 deletes the portion where the dividing line intersects.
Thereby, even if the small images are not properly arranged in a grid pattern, it is possible to generate a dividing line for accurately dividing the small image.

  FIG. 21 is an operation flow chart of the background lightness detection in the step 31. This flow includes a step 41 for obtaining the lightness frequency distribution of the images constituting the image captured by the image capturing unit 2 by the background detection means 4. Step 42 for obtaining an average value of the brightness frequency, and Step 43 for setting the brightness having a frequency higher than the average value of the brightness frequency as the background brightness. As shown in the frequency distribution of lightness of the pixels in FIG. 8, when the background lightness is dark, a peak exists in a dark part of the lightness, and there is a frequency distribution in which there is no particular peak other than the dark part of the lightness with the peak. There are many. This is because the background lightness is high because the background lightness is the same, while the image portion often has a wide variety of lightnesses, so the frequency does not increase compared to the background lightness.

Next, in step 42, the background detection means 4 obtains the average value 31 of the frequencies. The background detection means 4 sets the lightness having a frequency exceeding the average value of the frequencies as the background lightness.
The background lightness is the same not only in the case of darkness but also in the case of lightness or arbitrary lightness. The inventor detects the background lightness accurately and easily from the lightness having a frequency exceeding the average value of the image lightness. It became possible.

  In the case where the image area has no frame portion and is all images, the background detection unit 4 cannot detect the background lightness. In this case, in the lightness frequency distribution, the lightness frequency average value is calculated. Lightness having a frequency that exceeds the frequency is generally distributed in one place. In the case of such a lightness frequency distribution, the background detection means 4 can determine that there is no background.

  Regarding the detection of the divided image core portion in the step 34, as shown in FIG. 9, the image dividing means 5 obtains the frequency of the brightness other than the background brightness for each fine strip-shaped area parallel to the Y axis. This is represented as a histogram 41.

  Similarly, the image dividing unit 5 obtains the frequency of lightness other than the background lightness for each fine strip-shaped region parallel to the X axis, and represents this as a histogram 42.

  In the histograms 41 and 42 obtained in this way, the image dividing means 5 obtains a strip-shaped area having a frequency higher than a preset setting value, and determines the intersecting portion of the strip-shaped areas as a divided image core section. Detected as 43. This setting value is different from the setting value 44 for detecting the small image described above, and is a setting of the frequency of brightness that can detect minute information such as character information in addition to the small image.

Regarding the rectangle detection in which the core part of step 37 is enlarged, as shown in FIG. 10, the image dividing means 5 performs the four vertexes (A, B, C, D) of the divided image core part 43 in the image region. The rectangle ABCD constituting the divided image core unit 43 is extended in the outward direction until the background brightness is reached. In more detail, the vertex A is extended in the Y axis direction outside the rectangle ABCD until the lightness reaches the background lightness, and the point where the lightness becomes the background light is defined as Ay. Similarly, the vertex A is extended in the Y axis direction outside the rectangle ABCD until the lightness reaches the background lightness, and the point at which the background lightness is reached is defined as Ax. The same operation is performed for the remaining vertices B, C, and D.
The enlarged rectangle 51 is detected as a rectangle that includes all of Ax, Ay, Bx, By, Cx, Cy, Dx, and Dy thus obtained and has the smallest area.

  When the respective divided image core parts 63 are enlarged in this way, the enlarged rectangles 51 may overlap each other. In that case, as shown in the enlarged explanatory view of the combined rectangle of FIG. 11, the image dividing means 5 detects the combined rectangle 61 that includes all the overlapped enlarged rectangles 51 and has the smallest area. is there.

  In the determination of the number of rectangles in step 38, the number of the enlarged rectangle 51 and the combined rectangle 61 finally obtained in this way is counted, and it is determined whether or not the number is one. If the number is determined to be one, the fact is displayed on the screen or the like and the process is terminated. If it is determined that the number is not one, in step 39, the image dividing means 5 A dividing line for dividing each small image area in the area is generated. The dividing line can be provided at the opposite side of the adjacent final rectangle and in the middle of the side. At this time, the image dividing unit 5 checks the generated dividing line. If the dividing line intersects with the final rectangle, the image dividing unit 5 deletes the portion where the dividing line intersects and ends the process. Thereby, even if the small image areas are not aligned in a lattice pattern, it is possible to generate a dividing line for accurately dividing the small image area.

  As described above, in the medium accuracy division mode in the image processing apparatus of the present invention, when the image is not decomposed in the above-described high accuracy division mode, the background lightness is detected from the captured image, and lightness other than this background lightness is detected. By obtaining the frequency distribution, it is possible to accurately detect the arrangement of the small image area accompanied by minute information such as character information, and to easily divide the small image area, generate a dividing line for dividing the small image area, and the like. It becomes possible to carry out accurately.

  As shown in FIG. 1, an image processing apparatus, a processing method, and a processing program of the present invention include an OS, an application built on hardware resources including a computer CPU, memory, auxiliary storage device, display, input device, and the like. Since it is realized by a database, a network system, etc., and image processing is specifically realized by using the above hardware resources, it falls under the technical idea using the laws of nature, It can be used in industries such as computer manufacturing.

DESCRIPTION OF SYMBOLS 1 Image processing apparatus 2 Image acquisition part 3 Image processing part 4 Background detection means 5 Image division means 11 Scan line 12 Scan line 13 Title part 14 All image area seeds 15 Scan line part 16 All image areas 17 Small image 18 Temporary division line 19 Temporary dividing line 20 Small image area 21 Character information 31 Frequency average value 32 Background brightness range 41 Histogram 42 Histogram 43 Divided image core 51 Enlarged rectangle 61 Combined rectangle

Claims (11)

An image capture unit for capturing images;
Background detection means for detecting a background portion other than an image in the area of the captured image, a plurality of small images having image information included in the captured image, and character information associated with the small image for each small image An image processing apparatus comprising: an image dividing means for dividing; and an image processing unit comprising:
The background detection means detects a background portion from the brightness frequency distribution of the captured image,
The image dividing unit detects the small image from the image area other than the background part detected by the background detection unit using a frequency distribution of brightness other than the background part,
In addition, a band-like region obtained by dividing the captured image into the vertical direction and the horizontal direction including the small image and the character information by using the frequency distribution of the brightness other than the background portion is divided into a vertical scan line and a horizontal scan. Detect as line,
Detecting an overlapped area of the vertical scan line and the horizontal scan line as an entire image area seed for extracting an entire image area including the plurality of small images;
Set a boundary that encloses the outermost periphery of the entire image region seed in a rectangle and set the inner side as the entire image region,
The entire image area is determined from the entire image area and the arrangement of the small images so that each small image accompanied by the character information in the entire image area has an area of an equal area and is arranged near the center. Set a temporary dividing line that divides the image horizontally and vertically,
Dividing the entire image region seed by the temporary dividing line, dividing each small image accompanied by the character information, and combining the small image and the accompanying character information to detect a small image region. An image processing apparatus.   The background detection means obtains a lightness frequency distribution of pixels constituting the captured image and an average value of the lightness frequencies, and the lightness having a frequency exceeding the average value is the lightness of the background portion. The image processing apparatus according to claim 1, wherein:   The image dividing means obtains a frequency distribution of brightness other than background brightness among the brightness of pixels constituting the captured image, and a preset set value in an image area of the captured image based on the frequency distribution The image processing apparatus according to claim 1, wherein the arrangement of the small images is obtained from a frequency of brightness exceeding.   The image according to claim 1, wherein the background detection unit obtains periodicity in all image region seeds and excludes a single title portion having no periodicity in all image region seeds. Processing equipment.   When the provisional dividing line passes through the entire image region seed, the image dividing unit determines that the interval between adjacent small image regions is narrow, and determines the total region of all image region seeds to the number of detected small images. The image processing apparatus according to claim 1, wherein the image processing apparatus is divided into equal image areas to form a small image area. An image capture unit for capturing images;
Background detection means for detecting a background portion other than an image in the area of the captured image, a plurality of small images having image information included in the captured image, and character information associated with the small image for each small image A method of processing an image using an image processing device comprising an image dividing means for dividing, and an image processing unit comprising:
The background detection means detects a background portion from the brightness frequency distribution of the captured image,
The image dividing unit detects the small image from the image area other than the background part detected by the background detection unit using a frequency distribution of brightness other than the background part,
A strip-like region obtained by dividing the captured image into the vertical direction and the horizontal direction including the small image and the character information using a frequency distribution of brightness other than the background portion is divided into a vertical scan line and a horizontal scan line. Detect as
Detecting an overlapped area of the vertical scan line and the horizontal scan line as an entire image area seed for extracting an entire image area including the plurality of small images;
Set a boundary that encloses the outermost periphery of the entire image region seed in a rectangle and set the inner side as the entire image region,
The entire image area is determined from the entire image area and the arrangement of the small images so that each small image accompanied by the character information in the entire image area has an area of an equal area and is arranged near the center. Set a temporary dividing line that divides the image horizontally and vertically,
Dividing the entire image region seed by the temporary dividing line, dividing each small image accompanied by the character information, and combining the small image and the accompanying character information to detect a small image region. An image processing method.   The background detection means obtains a lightness frequency distribution of pixels constituting the captured image and an average value of the lightness frequencies, and the lightness having a frequency exceeding the average value is the lightness of the background portion. The image processing method according to claim 6, wherein:   The image dividing means obtains a frequency distribution of brightness other than the background brightness among the brightness of pixels constituting the captured image, and a preset set value in the image area of the captured image based on the frequency distribution The image processing method according to claim 6, wherein the arrangement of the small images is obtained from a frequency of brightness exceeding.   9. The image according to claim 6, wherein the background detection means obtains periodicity in all image region seeds and excludes a single title portion having no periodicity in all image region seeds. Processing method.   When the temporary dividing line passes through the entire image area seed by the image dividing means, it is determined that the interval between adjacent small image areas is narrow, and the total area of all the image area seeds is set to the number of detected small images. The image processing method according to claim 6, wherein the image processing method is divided into equal image areas corresponding to the small image areas. An image processing program for dividing a plurality of small images having image information from captured images and character information accompanying the small images for each small image,
An image capture unit for capturing images;
Background detection means for detecting a background portion other than an image in the area of the captured image, a plurality of small images having image information included in the captured image, and character information associated with the small image for each small image In an image processing apparatus comprising an image dividing means for dividing, and an image processing unit comprising:
Detecting a background portion from a frequency distribution of brightness of the image captured by the background detection means;
Detecting the small image from the image area other than the background portion detected by the background detection unit using the frequency distribution of brightness other than the background portion by the image dividing unit;
A band-like region formed by dividing the image captured by the image dividing unit into the vertical direction and the horizontal direction including the small image and the character information using a frequency distribution of brightness other than the background portion is formed as a vertical scan line. And detecting as a horizontal scan line;
Detecting a region overlapped by overlapping the vertical scan line and the horizontal scan line by the image dividing means as an entire image region seed for extracting all image regions including the plurality of small images;
Setting a boundary that surrounds the outermost periphery of the entire image region seed in a rectangle by the image dividing means and setting the inner side as the entire image region;
From the entire image area and the arrangement of the small images, the image dividing means causes each small image accompanied by the character information in the entire image area to have an area with an equal area and be arranged near the center. Setting a temporary dividing line for equally dividing the entire image area vertically and horizontally;
The image dividing means divides the entire image region seed by the temporary dividing line, divides each small image accompanied by the character information, and synthesizes the small image and the accompanying character information to form a small image region. Detecting step;
An image processing program for executing