JP4350317B2 - Image processing apparatus and method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus and method for extracting a specific pattern when different patterns are handled in image processing.
[0002]
[Prior art]
Conventionally, a character recognition technology (OCR: Optical Charactar Reader) is used as a technology for converting characters and symbols on paper into electronic data. This character recognition technology basically recognizes a clean pattern with little noise or the like. For example, FIG. 5 shows an example of a pattern having a lot of noise or the like for the character recognition / weaving technique (hereinafter referred to as “character recognition”). In such a pattern, the probability of erroneous recognition increases. The point that erroneous recognition is likely to occur in this pattern will be described below.
(1) The outline of the character is jagged (not shown in the figure, but is jagged over the entire character pattern. This mainly occurs when the resolution at the time of printing or optical input is not sufficient.)
(2) White pixels are scattered within the character line. (3) There are many noises. (4) The noise is not limited to characters with a one-dot width. There is a filtering technique as a pre-processing technique for shaping the pattern. Each filter needs to be tuned according to the target pattern. For example, filtering for character recognition should not impair the basic shape of the pattern. Therefore, in order to minimize the influence on the pattern, for example, a 3 × 3 pixel is referred to, and among the nine pixels, whether the surrounding eight pixels are white or black and the central pixel is black or white Will be determined. A processing example of the center pixel is shown in FIG. This method of creating a new image by determining whether all the pixels of the pattern image are white or black is called filtering processing. FIG. 7 shows an example in which the character recognition filtering process described above is performed on the image of FIG.
[0003]
As another example of the filtering process, there is a black frame filtering process only for shaping a black frame. An example is shown in FIG. As shown in the figure, unlike the character recognition filter processing, the purpose of this filter is to connect line breaks up to three dots wide in the horizontal and vertical directions. In this example, there is no area surrounded by a black frame in the original image, but there is an area surrounded by black pixels after the black frame filter processing, and the area within the black frame can be specified. Thus, the treatment of the filtering technique varies depending on the pattern to be extracted.
[0004]
In the above character recognition, for example, there are cases where two or more different patterns of a black frame and a character, such as a character being written in a black frame, are mixed. In order to recognize a character, it is necessary to have only a character pattern, and for this purpose, only a character part must be extracted from a black frame. Currently, a white continuous region extraction method is used as a method of cutting out images such as characters from the inside of a black frame. As shown in FIG. 8, the white continuous region extraction method obtains a white pixel continuous region surrounded by black pixels, calculates the coordinates of a frame cut out from the maximum horizontal width and the maximum vertical width, This is the extraction method.
However, in the white continuous area extraction method, the black frame in the image needs to be completely connected in principle in order to successfully cut out. However, for example, when a black frame form is acquired as electronic data from a fax or the like, the black frame is often drowned and cut. When the black frame is cut off, the white continuous area extraction method cannot cut out characters.
[0005]
[Problems to be solved by the invention]
If the black frame is cut off, it is conceivable to perform the black frame filtering process described above, and then cut out the character portion by the white continuous region extraction method. An example is shown in FIG. First, the black frame filter is applied to the original image (a). Since the filtering device cannot determine where the black frame is and how far the character pattern is, the black frame filter process is performed on the entire image. An image (b) in FIG. 10 is an example in which the black frame filter processing is performed on the original image (a). Next, a white pixel continuous region is obtained by the white continuous region extraction method, and a character pattern region to be passed to the character recognition unit is cut out. However, in this case, since the black frame filter processing is performed on the entire image, the black frame filter affects the character pattern. Even if the character recognition filter is applied to the cut out pattern, the image (c) in FIG. As shown, the character pattern changes.
[0006]
Also, even if the character recognition filter is first performed on the entire image without using the black frame filter, the character recognition filter cannot connect the defect of the frame of 2 dots width or more, so the black frame defect remains, After that, even if the white continuous area extraction method is applied, characters in the frame cannot be cut out. As described above, if the frame extraction filter is performed with priority on frame extraction, the accuracy of character recognition deteriorates. If the character recognition filter is performed with priority on character recognition, the accuracy of frame extraction decreases. Resulting in.
The present invention has been made in view of such circumstances, and an object of the present invention is to provide an image processing apparatus and method capable of extracting a target pattern with high accuracy when there are mixed patterns handled differently in image processing. And
[0007]
[Means for Solving the Problems]
The present invention has been made to achieve the above object, and includes a copying unit that creates a copy image by copying an original image, a first processing unit that performs image processing on the copy image, and the first processing. And a second processing unit that processes the original image using information acquired from the copy image processed in the unit. When the image processing apparatus of the present invention performs a filtering process on an image in which different patterns are mixed, the extraction accuracy of a specific pattern is improved.
[0008]
The present invention also provides a copy unit that creates a copy image by copying an original image, a first processing unit that executes a black frame filter process on the copy image, and a black frame filter that is used in the first processing unit. A second continuous area extraction method is applied to the copy image that has been processed to calculate the coordinates inside the black frame, and the calculated coordinates are applied to the original image to extract an image within the black frame. An image processing apparatus comprising a processing unit is provided. With the image processing apparatus of the present invention, an image within a black frame can be extracted without detracting from the characteristics of the image within the black frame.
[0009]
Further, the present invention provides a first process of creating a copy image by copying an original image, a second process of image processing of the copy image, and a copy image image-processed by the first processing unit. And a third step of processing the original image using the acquired information. In the image processing method of the present invention, when filtering processing is performed on an image in which different patterns are mixed, the extraction accuracy of a specific pattern is improved.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram illustrating a configuration of an image processing apparatus that cuts out characters and the like in a black frame from a black frame. Reference numeral 1 denotes a control unit which controls each unit. Reference numeral 2 denotes a copying unit, which copies an image. Reference numeral 3 denotes a first image processing unit that performs black frame filtering processing. Reference numeral 4 denotes a second image processing unit, which performs processing for cutting out characters in a black frame using a white continuous region extraction method.
[0011]
Next, the operation of this embodiment will be described with reference to FIGS. The control unit 1 receives an original image that is an object of image processing from a scanner or the like. In the original image, for example, characters enclosed in a frame as shown in image a in FIG. 3 are drawn. The control unit 1 instructs the copying unit 2 to copy the original image. The copying unit 2 copies the original image and creates a copy image as shown in the image b in FIG. 3 (step S10 in FIG. 2, (1) in FIG. 3). The copying unit 2 returns to the control unit 1 that the original image has been copied to create a copy image.
Next, the control unit 1 instructs the first image processing unit 3 to perform image processing on the copy image. The first image processing unit 3 performs black frame filtering processing on the entire copy image, as shown in an example in FIG. 9, by connecting black framed portions with black pixels, up to a width of 3 dots in the horizontal and vertical directions. 2 is performed (step S11 in FIG. 2, (2) in FIG. 3). An image c in FIG. 3 is an example in which the black frame filtering process is performed on the image b in FIG. 3. The first image processing unit 3 returns to the control unit 1 that the image processing has been performed on the copy image.
[0012]
The control unit 1 instructs the second image processing unit 4 to perform image processing. The second image processing unit 4 obtains a white pixel continuous region surrounded by black pixels with respect to the image processed by the first image processing unit 3 using the white continuous region extraction method shown in FIG. The character frame coordinates to be cut out are calculated from the maximum horizontal width and maximum vertical width. Next, as shown in an example in FIG. 4, the cutout area is specified by adapting the calculated character frame coordinates to the original image ((3) in FIG. 3), and the characters inside the black frame are extracted (step in FIG. 2). S12, (4) in FIG. The second image processing unit 4 returns to the control unit 1 that image processing has been performed. The control unit 1 outputs the image extracted by the second image processing unit 4 to a character recognition unit or the like.
[0013]
As described above, the image processing apparatus creates a copy image by copying the original image to be processed to the copy unit 2. The image processing apparatus processes the original image copied to the copying unit 2 using information obtained as a result of processing of the copy image by several image processing units. In the present embodiment, the first image processing unit 3 performs black frame filtering processing, and the second image processing unit 4 performs processing by the white continuous region extraction method. However, the present invention is not limited to this. is not. Further, the number of image processing units is not limited to two. In short, copy the original image to be processed to create a copy image, perform image processing on the copy image (the same image as the original image), and process the original image using information obtained as a result of the image processing Anything is acceptable.
[0014]
A program for realizing part or all of the functions of the control unit 1, the copying unit 2, the first image processing unit 3, and the second image processing unit 4 in FIG. 1 is recorded on a computer-readable recording medium. Thus, the functions of the control unit 1, the copying unit 2, the first image processing unit 3, and the second image processing unit 4 can be realized by causing the computer system to read and execute the program recorded on the recording medium. Good. Here, the “computer system” includes an OS and hardware such as peripheral devices.
Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
[0015]
The “computer-readable recording medium” refers to a portable medium such as a floppy disk, a magneto-optical disk, a ROM, or a CD-ROM, and a storage device such as a hard disk built in the computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.
The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design changes and the like within a scope not departing from the gist of the present invention.
[0016]
【The invention's effect】
As described above, according to the image processing apparatus and method of the present invention, the copying unit that copies the original image to create a copy image, the first processing unit that performs image processing on the copy image, and the first And a second processing unit that processes the original image using information acquired from the copy image processed by the processing unit, so that filtering processing is performed on an image having different patterns In this case, the extraction accuracy of the specific pattern is improved.
[Brief description of the drawings]
FIG. 1 is a block diagram illustrating a configuration of an image processing apparatus that cuts out an image within a black frame from a black frame according to an embodiment of the present invention.
FIG. 2 is a flowchart of the operation in the same embodiment.
FIG. 3 is an explanatory diagram of an operation of cutting out an image pattern in a black frame from a black frame in the embodiment.
FIG. 4 is an explanatory diagram for adapting a cutout region obtained by a white continuous region extraction method to an original image.
FIG. 5 is an example of a character pattern with a lot of noise.
FIG. 6 is an explanatory diagram of character recognition filter processing;
FIG. 7 is an example in which a character recognition filter is executed on a character with a lot of noise and the like.
FIG. 8 is an explanatory diagram of a white continuous region extraction method.
FIG. 9 is an example in which a black frame filter is executed.
FIG. 10 is an example in which a black frame filter is performed on a character image surrounded by a frame, and then a character recognition filter is performed.
[Explanation of symbols]
1: Control unit 2: Copy unit 3: First image processing unit 4: Second image processing unit

Claims (2)

A copy section that creates a copy image by copying the original image;
A first processing unit that executes black frame filter processing for connecting a black frame line to the copy image;
The white continuous region extraction method is applied to the copy image that has been subjected to the black frame filtering process in the first processing unit to calculate the coordinates inside the black frame, and the calculated coordinates are applied to the original image. A second processing unit for extracting an image within a black frame;
An image processing apparatus comprising: A first process of copying the original image to create a copy image;
A second step of performing a black frame filtering process for connecting a black frame line to the copy image;
The white continuous region extraction method is applied to the copy image that has been subjected to the black frame filtering process in the second step to calculate the coordinates inside the black frame, and the calculated coordinates are applied to the original image to apply the black frame A third step of extracting the image in the
An image processing method comprising:

Images