JP2022027060A - Image analysis system, image analysis method and image analysis program - Google Patents

Abstract

To perform an image analysis more accurately.SOLUTION: An image analysis system S comprises: a table detection unit 212; a frame detection unit 213; and an information acquisition unit 215. The table detection unit 212 is configured to detect an area corresponding to a table from an image including the table as a subject. The frame detection unit 213 is configured to detect a plurality of frames from the area corresponding to the table, and give each of the plurality of detected frames position information in the image. The information acquisition unit 215 is configured to acquire acquisition object information on the basis of a character recognition result about each of the plurality of frames, and the position information given to each of the plurality of frames.SELECTED DRAWING: Figure 3

Description

Application for application of Article 30, Paragraph 2 of the Patent Act was announced on October 14, 1991 at IECON 2019-45th Annual Conference of the IEEE Electrical Electricals Society.

The present invention relates to an image analysis system, an image analysis method, and an image analysis program.

Conventionally, there is a technique for performing image analysis such as character recognition on a paper medium or the like on which characters and tables are written.
An example of a technique related to such image analysis is disclosed in Patent Document 1. In the technique disclosed in Patent Document 1, it is set in advance which information is described in which position corresponding to the format (that is, appearance) of the manuscript to be read. Then, based on this setting, it is possible to perform image analysis such as detecting a character string from a position suitable for the format in the image.

Japanese Unexamined Patent Publication No. 2019-57311

However, the formats of the documents to be read are not necessarily the same, and various formats may be mixed for each document. In this case, it is practically difficult to set the format in advance or modify the format each time it is read, as in the technique disclosed in Patent Document 1.

Further, in the general technique disclosed in Patent Document 1 and the like, the manuscript is placed on a manuscript stand such as a multifunction device or a mechanism such as an ADF (Auto Document Feeder), and the scanner and the manuscript are kept in an appropriate state at a predetermined distance. It is assumed that the image is generated by being read by. However, it is not always read in an appropriate state to generate an image. For example, when a user shoots a document with a portable camera and generates an image, the document is curved at the time of shooting, or the positional relationship between the camera and the document at the time of shooting is not appropriate, and the image is distorted. It can happen.
As described above, it is desired to perform image analysis with high accuracy even when various formats are mixed or distortion occurs in the image.

The present invention has been made in view of such a situation. An object of the present invention is to perform image analysis with higher accuracy.

In order to solve the above problems, the image analysis system according to the embodiment of the present invention is
A table detecting means for detecting an area corresponding to the table from an image including the table as a subject,
A frame detecting means for detecting a plurality of frames constituting the table from a region corresponding to the table and assigning position information in the image to each of the detected frames.
An information acquisition means for acquiring acquisition target information based on the character recognition result for each of the plurality of frames and the position information given to each of the plurality of frames.
It is characterized by having.

According to the present invention, it is possible to perform image analysis with higher accuracy.

It is a block diagram which shows an example of the whole structure of the image analysis system which concerns on one Embodiment of this invention. It is a block diagram which shows an example of the structure of the terminal which concerns on one Embodiment of this invention. It is a block diagram which shows an example of the structure of the image analysis apparatus which concerns on one Embodiment of this invention. It is a flowchart explaining the flow of a shooting process. It is a flowchart explaining the flow of image analysis processing. It is a schematic diagram which shows that the frame line of a 1st direction and the frame line of a 2nd direction are separated and detected. It is a schematic diagram which shows the table structure included in the target image, and the acquisition method of the acquisition target information. It is a flowchart explaining the flow of a character detection process. It is a schematic diagram which shows that the out-of-frame area is deleted from the binarized image, and then the projective transformation is performed.

Hereinafter, an example of an embodiment of the present invention will be described with reference to the accompanying drawings.

[System configuration]
FIG. 1 is a block diagram showing an overall configuration of an image analysis system S according to the present embodiment. As shown in FIG. 1, the image analysis system S includes a terminal 1 and an image analysis device 2. Further, FIG. 1 also shows a manuscript 3 to be subjected to image analysis processing in the present embodiment. Here, it is assumed that the manuscript 3 is a manuscript including some kind of table.

These terminals 1 and the image analysis device 2 are connected to each other so as to be able to communicate with each other directly or via a network (not shown). In this case, the network is realized by, for example, a network such as the Internet, a LAN (Local Area Network), a mobile phone network, or a network in which these are combined. Further, this communication may be performed in accordance with any communication method, and the communication method is not particularly limited. Further, the connection format of communication may be a wired connection or a wireless connection.

The terminal 1 is a terminal having a shooting function. The terminal 1 is realized by, for example, a smartphone, a tablet-type terminal, a digital camera, or the like. The terminal 1 performs a shooting process. Here, the photographing process is a series of processes for generating an image (hereinafter, referred to as "target image") in which the document 3 is included as a subject by photographing the document 3. In the shooting process, the terminal 1 shoots based on the user's operation to generate a target image in which the document 3 is included as a subject. Then, the terminal 1 transmits the generated target image to the image analysis device 2.

The image analysis device 2 is a device that performs image analysis processing. The image analysis device 2 is realized by, for example, a personal computer, a server device, or the like. The image analysis device 2 acquires the target image transmitted from the terminal 1 by receiving it. Then, the image analysis device 2 performs an image analysis process on the target image. Here, the image analysis process is a series of processes for performing image analysis such as character recognition for detecting characters described in a target image.

In this image analysis process, the image analysis device 2 detects a region corresponding to the table from the image including the table (for example, the table) as a subject. Further, the image analysis device 2 detects a plurality of frames (for example, cells) constituting the table from the area corresponding to the table, and imparts position information in the image to each of the detected frames. Further, the image analysis device 2 has the acquisition target information (for example, the inspection item and the inspection value in the health diagnosis table) based on the character recognition result for each of the plurality of frames and the position information given to each of the plurality of frames. Pair) is acquired.

As described above, the image analysis system S including the terminal 1 and the image analysis device 2 sequentially detects the table and the frame, and then uses the position information to acquire the desired acquisition target information. Therefore, it is possible to perform image analysis with higher accuracy than an image analysis method in which the image analysis process is simply performed without detecting the table and the frame and without using the position information. Therefore, for example, when various formats are mixed and the arrangement of tables and frames differs depending on the original, or the image is distorted, it is usually difficult to perform image analysis. Even in some cases, image analysis can be performed appropriately.

Next, each device included in the image analysis system S for realizing such an image analysis process will be described in more detail.

In the following, as an example for explanation, it is assumed that the manuscript 3 is a “health diagnosis table” for a user who uses the image analysis system S. The reason will be explained.
First of all, health examinations are very useful for early detection and treatment of illness. In addition, by regularly undergoing a health examination, the user can grasp changes in his / her health condition and the like.

However, the format (that is, the appearance) of the health examination table differs depending on the business operator that creates it and the inspection executing agency. For example, the arrangement of tables and frames differs depending on the health examination table. In addition, the description position of the test item, the description position of the corresponding test value, and the description position of the reference value and the previous test value differ depending on the health diagnosis table. Therefore, it is difficult to perform image analysis of the health diagnosis table with general techniques.
In addition, electronic data of the health checkup table is not shared among the inspection executing agencies. Further, the user who is the examinee only receives the health examination table printed on the paper medium from each examination executing agency, and does not receive the electronic data in the first place. Under such circumstances, it is difficult for the user to grasp changes in his / her health condition. In addition, it is difficult for each inspection executing agency, hospital, etc. to grasp changes in the user's health condition.

Therefore, in the present embodiment, the above-mentioned image analysis processing is performed on the target image taken with such a health diagnosis table as a subject. Thereby, in the present embodiment, the image analysis can be performed with higher accuracy from the health diagnosis table. Further, as a result, the user can convert the health diagnosis table owned by the user into electronic data by a simple operation such as taking a picture with the terminal 1. Further, by using such electronic data, for example, it is possible for the user to easily grasp the change in the health state along with his / her own time series. In addition, such electronic data can be utilized in each inspection executing agency, hospital, or the like.
As described above, the health diagnosis table is suitable as a target for image analysis processing in the present embodiment. Therefore, in the following description, it is assumed that the manuscript 3 is a health diagnosis table.

However, this is merely an example for explanation, and does not limit the use of this embodiment. That is, the present embodiment can be applied to various cases such that the manuscript 3 includes some kind of table other than the case where the manuscript 3 which is the subject of the target image is a health diagnosis table.

[Terminal configuration]
The configuration of the terminal 1 will be described with reference to FIG. FIG. 2 is a block diagram showing an example of the configuration of the terminal 1. As described above, the terminal 1 is realized by, for example, a smartphone, a tablet-type terminal, a digital camera, or the like.

As shown in FIG. 2, the terminal 1 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a communication unit 14, a drive 15, and a storage unit 16. , An input unit 17, an output unit 18, and an image pickup unit 19. Each of these parts is connected by a signal line and sends and receives signals to and from each other.

The CPU 11 executes various processes (for example, a process of acquiring a target image by photographing a document 3 as a subject) according to a program recorded in the ROM 12 or a program loaded from the storage unit 16 into the RAM 13.
Data and the like necessary for the CPU 11 to execute various processes are also appropriately stored in the RAM 13.

The communication unit 14 controls communication for the CPU 11 to communicate with another device (for example, the image analysis device 2).

A removable medium (not shown) made of a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is appropriately mounted on the drive 15. The program read from the removable media by the drive 15 is installed in the storage unit 16 as needed. Further, various data read from the removable media by the drive 15 are used for arithmetic processing in the CPU 11 as needed.

The storage unit 16 is composed of a storage device such as an HDD (hard disk drive) or SSD (Solid State Drive), and stores various data.
The input unit 17 is composed of various buttons and a touch panel, or an external input device such as a mouse and a keyboard, and inputs various information according to a user's instruction operation.
The output unit 18 is composed of a display, a speaker, or the like, and outputs images and sounds.

The image pickup unit 19 is composed of a lens, an image pickup element, and the like, and photographs a subject (for example, a document 3). By this shooting, the image pickup unit 19 generates an image (for example, a target image).

In the terminal 1, various processes in the present embodiment are executed by the cooperation of each of these parts. When various processes in this embodiment are executed, as shown in FIG. 3, the photographing control unit 111 and the terminal side notification unit 112 function in the CPU 11.
Further, an image storage unit 161 and an analysis result storage unit 162 are set in one area of the storage unit 16.
Data necessary for realizing processing is appropriately transmitted and received between these functional blocks, even if not specifically mentioned below.

The shooting control unit 111 controls shooting at the terminal 1. Therefore, the shooting control unit 111 displays a user interface for the user to shoot the document 3 as a subject, a live view image acquired by the imaging unit 19, and the like on the display included in the output unit 18. Further, the shooting control unit 111 generates and acquires an image by controlling the imaging unit 19 based on a shooting instruction operation from the user. As the image to be acquired, as described above, it is assumed that the target image includes the manuscript 3 (here, the health diagnosis table) as the subject.

The photographing control unit 111 stores the acquired target image in the image storage unit 161. That is, the image storage unit 161 functions as a storage unit for storing the target image.
Further, the photographing control unit 111 transmits the acquired target image to the image analysis device 2.

The terminal-side notification unit 112 acquires the image analysis result in the image analysis process for the target image transmitted from the image analysis device 2. Further, the terminal-side notification unit 112 stores the acquired image analysis result in the analysis result storage unit 162. That is, the analysis result storage unit 162 functions as a storage unit for storing the image analysis result.

Further, the terminal-side notification unit 112 notifies the user of the acquired image analysis result. By referring to the notified image analysis result, the user can grasp the content of the image analysis result (for example, the test value in his / her health diagnosis table). This notification to the user can be realized, for example, by displaying on a display included in the output unit 18, printing on a paper medium from a printer (not shown), or the like.

The specific content to be notified as the image analysis result will be described later together with the explanation of the operation with reference to the flowchart of FIG.

[Configuration of image analysis device]
Next, the configuration of the image analysis device 2 will be described with reference to FIG. FIG. 3 is a block diagram showing an example of the configuration of the image analysis device 2. As described above, the image analysis device 2 is realized by, for example, a personal computer, a server device, or the like.

As shown in FIG. 3, the image analysis device 2 includes a CPU 21, a ROM 22, a RAM 23, a communication unit 24, a drive 25, a storage unit 26, an input unit 27, and an output unit 28. .. Each of these parts is connected by a signal line and sends and receives signals to and from each other. The functions of these parts as hardware are the same as those of the same name provided in the terminal 1 described above with reference to FIG. Therefore, the duplicated description of the function as hardware will be omitted.

In the image analysis device 2, various processes in the present embodiment are executed by the cooperation of these parts. When various processes in this embodiment are executed, as shown in FIG. 3, in the CPU 21, the image acquisition unit 211, the table detection unit 212, the frame detection unit 213, the character recognition unit 214, and information acquisition are performed. The unit 215 and the device-side notification unit 216 function.
Further, an image storage unit 261 and an analysis result storage unit 263 are set in one area of the storage unit 26.
Data necessary for realizing processing is appropriately transmitted and received between these functional blocks, even if not specifically mentioned below.

The image acquisition unit 211 acquires the target image transmitted from the photographing control unit 111 included in the terminal 1 by receiving the target image. Further, the image acquisition unit 211 stores the acquired target image in the image storage unit 261. That is, the image storage unit 261 functions as a storage unit for storing the target image.

The photographing control unit 111 included in the terminal 1 may store the target image in the removable media via the drive 15. Then, the image acquisition unit 211 may acquire the target image by reading the removable media in which the target image is stored via the drive 25.

The table detection unit 212 is referred to as an area (hereinafter, referred to as “table area”) corresponding to each of a plurality of (may be singular) tables included in the target image from the target image stored in the image storage unit 261. ) Is detected. For example, the frame detection unit 213 detects a table area corresponding to each of a plurality of tables constituting the health diagnosis table that is the subject in the target image.

The frame detection unit 213 detects a plurality of frames constituting the table from each of the table areas detected by the table detection unit 212. For example, the frame detection unit 213 detects a plurality of frames constituting the health diagnosis table that is the subject in the target image. In this case, for example, the frame detection unit 213 separates and detects the frame line in the first direction and the frame line in the second direction, which are the frame lines of a plurality of frames constituting the table, from each of the table areas. , A plurality of frames are detected based on the intersection of the detected borders in the first direction and the borders in the second direction. In this way, the frame detection unit 213 separates and detects the frame line in the first direction and the frame line in the second direction, so that the frame line and the frame line can be detected without separating them. The frame can be detected accurately.

Further, the frame detection unit 213 divides and cuts out each of the detected plurality of frames, and imparts position information in the target image to each of the cut out plurality of frames. For example, the frame detection unit 213 uses a combination of the table identification information (for example, the identifier assigned to the table) assigned to each of the detected table areas and the frame identification information (for example, the identifier assigned to the frame) as position information. And this position information is given to each of a plurality of frames. Alternatively, for example, the frame detection unit 213 uses coordinate values in the image coordinate system used in the target image (for example, values in each coordinate axis corresponding to the center of the frame, etc.) as position information, and a plurality of the position information. It is given to each of the frames.
As a result, the frame detection unit 213 generates an image in which each of the plurality of frames included in the target image is a unit and to which position information is added (hereinafter, referred to as "frame image"). do. The frame image is, for example, a unit in which a specific character string (for example, test item, test value, reference value, previous test value, attribute information such as patient's name, etc.) is described in the case of a health diagnosis table. Yes, it can be handled as structured data with location information.

The character recognition unit 214 performs character recognition processing for each of the plurality of frame images generated by the frame detection unit 213. This character recognition process can be realized, for example, by using an existing optical character recognition (OCR) technique. In this way, the character recognition unit 214 can perform image analysis processing in units of frames including a specific character string based on the frame image generated by the frame detection unit 213, so that the entire image can be imaged. Character recognition can be performed with higher accuracy than when analysis processing is performed.
The character recognition unit 214 not only uses the existing optical character recognition technology, but also uses, for example, a character string that can be included in the target image in advance (here, a character string such as a word used in a health examination table). Machine learning may be performed. Then, the character recognition process may be performed using the learning model generated by this machine learning.

The information acquisition unit 215 acquires acquisition target information based on the character recognition result for each of the plurality of frame images by the character recognition unit 214 and the position information given to each of the plurality of frame images. Here, the acquisition target information is information desired to be acquired by the user, and is, for example, a set of "inspection items" and "inspection values" in the health diagnosis table.

In order to acquire the acquisition target information by the information acquisition unit 215, in the present embodiment, dictionary data for the first information which is a part of the acquisition target information is prepared. For example, when this first information is a "search item", dictionary data is prepared for the text of a word (for example, height, weight, blood pressure, etc.) that is expected to be used as a test item name. Here, the inspection item names in the health examination have fluctuations in the notation and refer to the same inspection items in terms of meaning, but the inspection item names may differ depending on the format of the health examination table. For example, even if the same test item is high density lipoprotein (HDL), the test item name in the health checklist may be "HDL cholesterol" or "HDL-C". It may be "HDL". Therefore, in consideration of the fluctuation of these notations, the dictionary data should be associated with the texts of a plurality of inspection item names even for the same inspection item.

Such dictionary data is, for example, created in advance by the administrator of the image analysis system S and stored in the dictionary data storage unit 262. That is, the dictionary data storage unit 262 functions as a storage unit for storing dictionary data.

Then, the information acquisition unit 215 stores the character recognition result (that is, the text of the character string included in each frame image) for each of the plurality of frame images by the character recognition unit 214 and the dictionary stored in the dictionary data storage unit 262. Score the degree of similarity with each text of the word used as the inspection item name included in the data. Further, as a result of the scoring, the information acquisition unit 215 uses the frame image containing the character string having the highest degree of similarity for the certain inspection item name as the certain inspection item (that is, the first information). Is marked as a frame image in which. For example, when a certain frame image contains the character string "HDL-C", the certain frame image is marked as a frame image in which the inspection item "high specific gravity lipoprotein" is described.

Next, the information acquisition unit 215 is a part of the acquisition target information based on the position information given to the marked frame image (that is, the frame image from which the first information is acquired). Get the information of 2. For example, assume that this second information is a "search value" for a certain inspection item. In this case, the information acquisition unit 215 acquires the position information of the frame image marked corresponding to this certain inspection item. Further, the information acquisition unit 215 scans the position information to check whether the inspection value is included in each frame image existing in an arbitrary predetermined direction (for example, to the right).

Here, the realistic upper limit value and lower limit value of the test value for each test item can be predicted from a medical point of view. Therefore, for example, in the dictionary data, a realistic upper limit value and lower limit value are associated with each inspection item. Then, in the process of performing a scanning search for a certain inspection item, the information acquisition unit 215 includes a value equal to or less than the upper limit value associated with the certain inspection item and more than or equal to the lower limit value. When the frame image is detected, it is specified that the frame image is a frame image in which the inspection value for this certain inspection item is described. Then, this inspection value is acquired as the second information. As a result, even in a table such as a health diagnosis table in which the relative positional relationship between the frame in which the first information is described and the frame in which the second information is described differs depending on the format. Since the scanning search is performed, it is possible to perform image analysis with high accuracy. In addition, since the second information is acquired based on the realistic upper limit value and lower limit value, it is possible to prevent the acquisition of erroneous second information, and image analysis can be performed with high accuracy from this viewpoint as well. It will be possible to do.

In this way, the information acquisition unit 215 receives the first information (for example, a certain inspection item) which is the acquisition target information and the second information corresponding to the first information (inspection value of the certain inspection item). You can get a pair. In addition, such marking and scanning search based on the position information of the marked frame image may be performed each time one first information (for example, a certain inspection item) is marked. Alternatively, when there are a plurality of first information, marking may be performed on all the first information, and then a scanning search may be performed.

The information acquisition unit 215 stores the acquisition target information acquired in this way (for example, a set of the inspection item name corresponding to the inspection item and the inspection value) as an image analysis result in the analysis result storage unit 263. That is, the analysis result storage unit 263 functions as a storage unit for storing the image analysis result. Further, the information acquisition unit 215 transmits the image analysis result to the terminal 1 so that the terminal 1 can also notify the user of the image analysis result.

The device-side notification unit 216 notifies the user of the image analysis result stored in the analysis result storage unit 263 by the information acquisition unit 215. The user can grasp the content of the image analysis result by referring to the notified image analysis result. The notification to this user is performed in the same manner as the terminal side notification unit 112 of the terminal 1. For example, it can be realized by displaying on a display included in the output unit 28, printing on a paper medium from a printer (not shown), or the like.

The specific content to be notified as the image analysis result will be described later together with the explanation of the operation with reference to the flowchart of FIG.

The configuration of each device included in the image analysis system S has been described in detail above. Next, the processing contents of each processing performed by each apparatus included in the image analysis system S will be described in more detail.

[Image analysis processing]
FIG. 4 is a flowchart illustrating a flow of shooting processing performed by the terminal 1. The shooting process is executed when the user turns on the power to the terminal 1.

In step S11, the photographing control unit 111 starts controlling the photographing in the terminal 1. For example, the shooting control unit 111 displays a user interface for the user to shoot with the document 3 as a subject, a live view image acquired by the imaging unit 19, and the like on a display included in the output unit 18.

In step S12, the shooting control unit 111 determines whether or not the shooting instruction operation from the user has been accepted. When the shooting instruction operation is accepted, it is determined as Yes in step S11, and the process proceeds to step S12. On the other hand, if the shooting instruction operation is not accepted, it is determined as No in step S11, and the process repeats the determination in step S12.

In step S13, the shooting control unit 111 generates and acquires the target image by controlling the imaging unit 19.

In step S14, the photographing control unit 111 transmits the target image acquired in step S13 to the image analysis device 2. This ends this process.
By the shooting process described above, the image analysis device 2 can acquire the target image.

[Image analysis processing]
FIG. 5 is a flowchart illustrating a flow of image analysis processing performed by the image analysis device 2. The image analysis process is executed when the target image is transmitted from the terminal 1 or when the user receives an instruction operation for starting the image analysis process to the image analysis image analysis device 2.

In step S21, the image acquisition unit 211 acquires the target image.

In step S22, the table detection unit 212 detects a table area corresponding to each of a plurality of (may be singular) tables included in the target image acquired in step S21.
Specifically, the table detection unit 212 first performs preprocessing. For example, the table detection unit 212 converts the target image into gray scale and adjusts (that is, resizes) the size to 1280 × 960. Next, the table detection unit 212 separates the background area, the frame line, and the character area in the target image, and performs binarization for expressing by 0 or 1. For example, the table detection unit 212 performs binarization using retinex filtering. In binarization, for example, in an image, white is represented by 0 and black is represented by 1. Further, the table detection unit 212 performs filtering for deleting the character area. As a premise, the border of the table extends over a wide area in the target image, and characters are included in the border. Therefore, it is considered that the pixels are loosely connected. From this, it is considered that the variances in the x-direction and the y-direction in the frame line are both large.

Subsequently, the line segment in which the frame line of the table is divided by noise is considered to be extremely long in either the x direction or the y direction and extremely short in the other direction. From this, it is considered that the variance in the line segment is large in the x direction and the y direction and small in the other direction. Based on this property, for example, the table detection unit 212 identifies a line segment and a border by threshold processing using the eigenvalues of the covariance matrix in the x-direction and the y-direction of the image, and any of these line segments and borders. Areas that do not apply to are regarded as character areas and removed. Further, the table detection unit 212 performs thinning so that only one pixel at the center of the line segment and the frame line remains.

Then, the table detection unit 212 detects a region table area corresponding to each of a plurality of (sometimes singular) tables included in the target image from the target image. For example, when a plurality of tables exist in the target image, the table detection unit 212 divides the image so that only one table exists in one image. Therefore, the table detection unit 212 creates a cumulative energy map in the x-direction and the y-direction used in seam carving. Further, the table detection unit 212 performs backtracking in the valley portion composed of pixels whose energy difference from the adjacent pixels is equal to or larger than the threshold value in the energy map. Then, each table area is detected by using the pixels traced by this back tracking as the boundary between the table and the document area, and each table area is divided into an image corresponding to one table area.

In step S23, the frame detection unit 213 detects a plurality of frames constituting the table from each of the table areas detected in step S22.
Specifically, the frame detection unit 213 first separates the frame lines. In this case, as described above, the frame detection unit 213 separately detects the frame line in the first direction and the frame line in the second direction, which are the frame lines of the plurality of frames constituting the table. This point will be described with reference to FIG. FIG. 6 is a schematic diagram showing that the frame line in the first direction and the frame line in the second direction are separately detected. FIG. 6A shows a frame line before separation in the image. The borders of the plurality of frames constituting the table in this way include borders in the first direction (here, the vertical direction) and the second direction (here, the horizontal direction). The frame detection unit 213 arranges these frame lines in the first direction (here, the vertical direction) as shown in FIG. 6 (b-1) and the second direction as shown in FIG. 6 (b-2). It is detected separately from the border in the direction (here, the horizontal direction).

As a premise, the width of the frame line is adjusted to 1 by the thinning performed in step S22. Therefore, by deleting the pixel having a width of 1 in the first direction (here, the vertical direction) with respect to the image, the frame line in the second direction (here, the horizontal direction) is deleted. As a result, as shown in FIG. 6 (b-1), an image of only the border in the first direction (here, the vertical direction) can be obtained. Similarly, by deleting the pixel having a width of 1 in the second direction (here, the horizontal direction), the border in the first direction (here, the vertical direction) is deleted. As a result, as shown in FIG. 6 (b-2), an image of only the border in the second direction (here, the horizontal direction) can be obtained. As a result, the border in the first direction (here, the vertical direction) and the border in the second direction (here, the horizontal direction) can be separated, and the border can be detected accurately.

A case where such separation is not performed will be described. As a premise, when a paper medium such as a health examination table is photographed, distortion occurs in each of the vertical and horizontal borders due to the turning of the paper or the like. Therefore, if the separation is not performed, the accuracy of detecting the frame line will be lowered. On the other hand, in the present embodiment, the frame lines in a plurality of directions are separated by performing detection after separating the frame lines in the vertical direction and the horizontal direction. Then, each of the separated borders in a plurality of directions is detected as a curve with a limitation on the curvature of the line, which simplifies the detection of the borders in the plurality of directions and makes it possible to detect the borders in both the vertical and horizontal directions. The border can be detected accurately according to each distortion.

Next, the frame detection unit 213 detects the frame line. Since the borders in the plurality of directions are separated as described above, the separated images are the borders in the first direction (here, the vertical direction) or the second direction (here, the horizontal direction), respectively. Direction) only. Therefore, the set of pixels obtained by looking at the connection relationship of the vicinity of 8 with respect to these images becomes a frame line in the vertical direction or the horizontal direction, respectively. In this embodiment, the curve approximation is performed by using the least squares method for this set of pixels.

Specifically, a set of pixels is detected by looking at the connection relationship in the vicinity of 8 with respect to the separated image. In addition to the border of the table, the set of pixels obtained by this includes characters that were not deleted because they were adjacent to the border during filtering. Therefore, the characters are deleted by the threshold processing using the number of pixels of the pixel set used in the filtering and the eigenvalues of the covariance matrix.
In addition, curve approximation is performed from a set of pixels with a large number of pixels, the order is gradually increased from 1 to 4 using a polynomial, the threshold value is determined as shown in the following equation (1), and only when the value is equal to or less than the threshold value. Create an approximate curve.

ただし、式（１）において、ｘ及びｙは座標である。

However, in the equation (1), x and y are coordinates.

Here, the threshold value (for example, 1.1 described above in the equation (1)) can be appropriately set. Then, when the approximate curve is created, the separated borders are combined. To synthesize the frame line, search the pixel set for the three pixels around the created approximate curve. When a pixel set is found, an iterative method of performing curve approximation by the least squares method is used again using both the pixel set and the pixel set used for curve approximation.

As soon as the approximation of the curve for all the pixel sets is completed, the frame detection unit 213 deletes the intersection line. The borders of the table are composed of parallel lines, and vertical lines and horizontal lines do not intersect each other. Therefore, if the lines intersect in the frame line image separated vertically and horizontally, it is highly likely that the lines outside the frame of the table are mixed. In the present embodiment, when the vertical and horizontal frame lines intersect, the out-of-frame line is deleted by deleting the intersecting lines having a small number of pixels in the pixel set used for approximation.

Then, the frame detection unit 213 detects the frame. In frame detection, the frame structure is detected by lines. When the above-mentioned border detection is completed, a group of vertical and horizontal approximate curves is generated. In the frame detection in this step, first, the position of the intersection of the approximate curves in the vertical direction and the horizontal direction is specified. The coordinates of the intersections obtained in this way become candidate positions for the corners of the frame, and as many as the number of intersections for the number of horizontal borders can be obtained for one vertical border, and for one horizontal border. You can get the number of intersections for the number of vertical borders.

After that, one vertical or horizontal border is selected, and one pair of adjacent intersections existing on the approximate curve is selected. Then, by searching between the intersections in the border image and determining that it is a border when a certain number of pixels exist, this process is applied to all pairs of adjacent intersections on the approximate curve in the image. Borders can be detected. After that, the area surrounded by the detected frame line is regarded as a frame and detected.

Further, the frame detection unit 213 divides and cuts out each of the plurality of detected frames. As described above, the frame detection unit 213 detects the frame based on the frame line surrounding the frame, and is therefore the outermost in the x and y coordinates among the coordinates included in the frame line surrounding the frame. The coordinates of are searched, and each of a plurality of frames is divided and cut out using the coordinates obtained by this search. Further, the frame detection unit 213 imparts position information in the target image to each of the plurality of cut out frames. For example, the frame detection unit 213 uses a combination of the table identification information (for example, the identifier assigned to the table) assigned to each of the detected table areas and the frame identification information (for example, the identifier assigned to the frame) as position information. And this position information is given to each of a plurality of frames. Alternatively, for example, the frame detection unit 213 uses coordinate values in the image coordinate system used in the target image (for example, values in each coordinate axis corresponding to the center of the frame, etc.) as position information, and a plurality of the position information. It is given to each of the frames. As a result, the frame detection unit 213 generates a frame image in which each of the plurality of frames included in the target image is a unit and a position information is added to each of the frames.

Returning to FIG. 5, in step S24, the character recognition unit 214 performs character recognition processing using the frame image as a unit for each of the plurality of frame images generated in step S23.

In step S25, the information acquisition unit 215 acquires acquisition target information based on the character recognition result for each of the plurality of frame images recognized in step S24 and the position information given to each of the plurality of frame images. To get. This point will be described with reference to FIG. 7. FIG. 7 is a schematic diagram showing a table structure included in the target image and a method of acquiring the acquisition target information by the information acquisition unit 215.

FIG. 7A shows a table structure included in the target image. In FIG. 7A, the target image 5 includes a table 51 and a plurality of frames 52 (in the figure, only one frame is designated for convenience of illustration). Although only one table 51 is shown in the figure for the sake of brevity, the target image 5 may include a plurality of tables 51. By performing the above-mentioned image analysis processing on such a target image 5, the table detection unit 212 detects the table area corresponding to the table 51. Further, the frame detection unit 213 generates a frame image for each of the plurality of frames 52 included in the table 51.

FIG. 7B shows a method of acquiring information to be acquired by the information acquisition unit 215. As described above, the information acquisition unit 215 stores the character recognition result (that is, the text of the character string included in each frame image) for each of the plurality of frame images by the character recognition unit 214 and the dictionary data storage unit 262. The degree of similarity with each text of the word used as the inspection item name included in the dictionary data is scored. Further, the information acquisition unit 215 assumes that, as a result of the scoring, a frame image containing the character string having the highest degree of similarity for a certain inspection item name is a frame image in which the inspection item name is described. Mark. For example, assume that the frame 52a is marked this time. Then, the information acquisition unit 215 acquires the position information of the frame image 52a marked corresponding to this certain inspection item. Further, as shown by an arrow in the figure, the information acquisition unit 215 includes inspection values for each frame image existing in an arbitrary predetermined direction (for example, to the right) from the position information of the frame image 52a. Search by scanning to see if it is.

In this way, the information acquisition unit 215 receives the first information (for example, a certain inspection item) which is the acquisition target information and the second information corresponding to the first information (inspection value of the certain inspection item). You can get a pair. The acquisition target information acquired in this way (for example, the inspection item name corresponding to the inspection item and the set of the inspection values) is used as the image analysis result.

In step S26, the terminal-side notification unit 112 or the device-side notification unit 216 notifies the user of the image analysis result in step S25. The user can grasp the content of the image analysis result by referring to the notified image analysis result. Note that the notification may be performed only by either the terminal-side notification unit 112 or the device-side notification unit 216, or may be performed by both. This ends this process.

Here, the notification in step S26 may be merely notified by simply displaying the image analysis result or the like, but the notification is not limited to this. For example, when image analysis processing is performed on a plurality of health examination tables corresponding to one user (for example, each of the health examination tables for each year when the health examination is performed every year). , The change in the time series of the inspection value corresponding to the predetermined inspection item may be detected based on the image analysis result, and the detected change may be notified.

In this case, for example, the implementation date described in each health diagnosis table is detected by character recognition, or the order in which each health diagnosis is performed is specified based on the user's operation or the like. Then, the result of the health diagnosis (for example, the test value) recognized by characters from each health diagnosis table is notified in correspondence with the order in which the health diagnosis is performed. For example, notification is given by displaying the test values of the same test items in the form of a table in association with the order in which the health examinations are performed. Alternatively, for example, the notification is given by displaying in the form of a graph in which the horizontal axis is the time (for example, the date when the health examination was performed) and the vertical axis is the test value of the same test item.

In this case, for example, a threshold value indicating an appropriate range of the inspection value (for example, an upper limit value or a lower limit value) is set, and when the inspection value is outside the appropriate range (for example, the upper limit value or more). Or, if it is below the lower limit value), the inspection value may be displayed in a different color or displayed together with a warning message to notify the user. In addition, for example, when a threshold value is set for the amount of change in the test value from the previous medical examination and the amount of change in the test value exceeds the threshold value (for example, when the test value changes significantly), The inspection value may be displayed in a different color or the like, or may be displayed together with a warning text to notify the user. As a result, the user can not only grasp the change of the inspection value of the same inspection item along the time series, but also grasp the change such that the inspection value is out of the appropriate range or has changed significantly. can.
Therefore, the user can more appropriately grasp the result of the health diagnosis.

According to the shooting process and the image analysis process described above, the table and the frame are detected step by step, and then the desired acquisition target information is acquired by using the position information. Therefore, it is possible to perform image analysis with higher accuracy than an image analysis method in which the image analysis process is simply performed without detecting the table and the frame and without using the position information. Therefore, for example, when various formats are mixed and the arrangement of tables and frames differs depending on the original, or the image is distorted, it is usually difficult to perform image analysis. Even in some cases, image analysis can be performed appropriately.

[Modification example]
Although the embodiment of the present invention has been described above, this embodiment is merely an example and does not limit the technical scope of the present invention. The present invention can take various other embodiments without departing from the gist of the present invention, and can be modified in various ways such as omission and substitution. In this case, these embodiments and variations thereof are included in the scope and gist of the invention described in the present specification and the like, and are included in the scope of the invention described in the claims and the equivalent scope thereof.
As an example, the embodiment of the present invention described above may be modified as follows.

In the above-described embodiment, the image analysis system S is realized as a client-server type system such as a combination of the terminal 1 and the image analysis device 2. Not limited to this, the image analysis system S may be configured with a configuration different from the example shown in the above-described embodiment.

For example, the image analysis device 2 may be realized not by one computer but by a plurality of computers. In this case, for example, a part of the functional blocks of the image analysis device 2 (for example, the character recognition unit 214 and the information acquisition unit 215 are realized by a certain computer, and the other functional blocks are realized by another computer. Further, for example, a part or all of the functional blocks of the image analysis device 2 may be realized by distributed processing by a plurality of computers (for example, a cloud system).

In addition, for example, by providing an image pickup unit in the image analysis device 2, the functions of the terminal 1 and the image analysis device 2 may be realized by one computer. In this case, the communication function may be omitted from this computer and the device may be used as a stand-alone device.

Furthermore, the terminal 1 may be used by the user who has undergone the health examination, or may be another entity such as a business operator who employs the user who has undergone the health examination. ..

Further, for example, in order to further improve the accuracy of the character recognition processing in step S24, the character detection processing may be performed on each of the plurality of frame images generated in step S23. This process is performed, for example, by the frame detection unit 213 that has generated a plurality of frame images, following the generation of the frame images. In this case, the frame detection unit 213 detects characters from the target image for each region, using a plurality of frame images detected by itself as a unit. As a result, the frame detection unit 213 selects a set of character strings (for example, one word) from a predetermined area (for example, in the frame in which characters are described) based on the accurate detection result of the frame by itself. It can be detected accurately.

FIG. 8 is a flowchart illustrating the flow of the character detection process.
In step S31, the frame detection unit 213 acquires the frame image to be processed this time from the plurality of frame images generated by itself.

In step S32, the frame detection unit 213 binarizes. In this modification, characters are detected by using the projection method in step S35 described later. When the projection method is used, it is necessary to perform binarization in order to count the black pixels in the character area and separate the character area and the background area. The binarization can be performed by any method, and for example, the binarization of Otsu is used. Otsu's binarization is also called a discriminant analysis method, in which a histogram of the brightness value is created for the image, and when the histogram is divided into two, the binarization is performed using the threshold value that maximizes the interclass variance. be. Many health checklists use various colors, but if you look inside the frame, it is a binary image of characters and background, so by using Otsu's binarization, the optimum threshold is automatically set. It is possible to calculate and extract characters.

In step S33, the frame detection unit 213 deletes the out-of-frame area. This point will be described with reference to FIG. FIG. 9 is a schematic diagram showing that the out-of-frame area is deleted from the binarized image, and then the projective transformation is performed. In step S32, the frame image was cut out using the outermost one from the coordinates of the four frame lines used to detect the frame. Therefore, when the frame line is inclined or distorted, characters outside the frame and the frame line are included and detected as shown as a binarized image in FIG. 9 (a). These characters and borders become obstacles in later character detection. Therefore, by deleting the pixels outside the frame line used in the frame detection, these can be deleted as shown in FIG. 9B.

In step S34, the frame detection unit 213 performs a projective transformation. The image of the target image acquired by the camera shooting causes distortion of the paper and distortion based on the positional relationship between the camera and the shooting target. A projective transformation is used to deal with this distortion. The conversion formula of the projective transformation is expressed by the following formula (2), and can be executed when H is regular.

ただし、式（２）において、ｘ’及びｙ’は変換後の座標であり、ｘ及びｙは変換前の座標であり、Ｈ_１１～Ｈ_３３は変換係数である。また、式（２）において、ｆ_０はスケールを表す定数であり、ｘ／ｆ_０及びｙ／ｆ_０が０になるように調整される。

However, in the equation (2), x'and y'are the coordinates after the conversion, x and y are the coordinates before the conversion, and H ₁₁ to H ₃₃ are the conversion coefficients. Further, in the equation (2), f ₀ is a constant representing a scale, and is adjusted so that x / f ₀ and y / f ₀ become 0.

Since the transformation matrix of this projective transformation contains eight transformation coefficients and two equations can be obtained from the coordinates before and after the transformation, the transformation coefficients are prepared by preparing four pairs of coordinates before and after the transformation. Can be asked. In this modification, the intersection of the frame lines is used for these four coordinates. For the coordinates after conversion, the coordinates of the upper left of the frame are fixed, the distance between the intersections of the upper left and upper right, and the lower left and lower right is calculated in the x-axis direction, and the longer distance is the length of the horizontal frame after conversion. Let's say. Similarly, the distance in the y-axis direction between the intersections of the upper left and lower left, and the intersections of upper right and lower right is calculated, and the longer distance is used as the length of the converted vertical frame, and the converted frame becomes a rectangle. Define the coordinates as follows. The coordinates of the intersection are used as they are for the coordinates before the conversion, the conversion coefficient is determined using these four coordinates, and the projective transformation is performed as shown in FIG. 9 (c).

In step S35, the frame detection unit 213 detects characters. A projection method is used for character detection. In the projection method, the number of pixels of the pixel corresponding to black is counted after binarization at each coordinate in each of the y and x directions of the image, and the range where the number of pixels is not 0 is detected as a line and a character. However, with this method, katakana components such as "ri" and "ru", and the bias and sword of kanji such as "ken" and "contact" are detected separately on the left and right. Problems arise. Therefore, in this modification, characters are combined after being separated and detected.

In this case, first, for the characters detected by the projection method, the horizontal length of each character is divided by the vertical length, as expressed by the following equation (3), and for each detected character. Calculate the aspect ratio.

ただし、式（３）において、ｒ_ｉは各文字の縦横比であり、ｗ_ｉは各文字の幅であり、ｈは行の高さである。

However, in the equation (3), r _i is the aspect ratio of each character, _wi is the width of each character, and h is the height of the line.

This aspect ratio _ri is relatively large in the case of full-width characters and relatively small in the case of half-width characters. The health diagnosis table, which is the target of image analysis processing, has a character area such as an inspection item and a numerical area of the inspection result, and the numerical area is described in half-width characters. Therefore, if the characters existing in the frame are only half-width characters, it is considered that character composition is not necessary. Therefore, the character area and the numerical area of the table are discriminated by using this aspect ratio _ri . Therefore, a threshold value for discrimination is set, and if a threshold value or more exists, it is regarded as a character area, and the characters in the frame of the character area are used as candidates for character composition.

Whether or not to actually synthesize characters for the candidates for synthesizing these characters is examined for adjacent characters that satisfy the conditions represented by the following equation (4).

ただし、式（４）において、ｘ_{ｉ＋１，１}及びｘ_ｉ，２は文字の横方向の始点及び終点であり、ｓ_ｉは検出文字間の距離である。

However, in the equation (4), x _{i + 1} , 1 and x _{i , 2} are the start and end points in the horizontal direction of the character, and si is the distance between the detected characters.

Then, it is examined whether or not the characters are combined in ascending order of the distance, and the aspect ratio when the characters are combined is calculated by using the above equation (3). Then, if the calculated aspect ratio is equal to or greater than the threshold value, the characters are actually combined as if they were single characters detected separately in the character area. However, if the calculated aspect ratio is less than the threshold value, it is not actually synthesized as a single character detected without separation in the character area. As a result, for example, when the numerical values described in half-width characters are not mistakenly combined, and on the other hand, the bias (hen) and the 旁 (making) of the Chinese characters are detected separately. Can be combined as a single character. This ends this process. Then, the character recognition unit 214 can perform the character recognition process in step S24 with high accuracy on a group of character strings (for example, one word) accurately detected by this process.

[Configuration example]
As described above, the image analysis system S according to the present embodiment includes a table detection unit 212, a frame detection unit 213, and an information acquisition unit 215.
The table detection unit 212 detects an area corresponding to the table from an image including the table as a subject.
The frame detection unit 213 detects a plurality of frames constituting the table from the area corresponding to the table, and imparts position information in the image to each of the detected plurality of frames.
The information acquisition unit 215 acquires the acquisition target information based on the character recognition result for each of the plurality of frames and the position information given to each of the plurality of frames.

In this way, the image analysis system S acquires the desired acquisition target information by using the position information after detecting the table and the frame in stages. Therefore, it is possible to perform image analysis with higher accuracy than an image analysis method in which the image analysis process is simply performed without detecting the table and the frame and without using the position information. Therefore, for example, when various formats are mixed and the arrangement of tables and frames differs depending on the original, or the image is distorted, it is usually difficult to perform image analysis. Even in some cases, image analysis can be performed appropriately.

The information acquisition unit 215 acquires the first information based on the dictionary data about the first information which is a part of the acquisition target information and the character recognition result of each of the plurality of frames.
The information acquisition unit 215 acquires the second information which is a part of the acquisition target information based on the position information corresponding to the frame from which the first information is acquired.
As a result, image analysis can be performed with higher accuracy by using the dictionary data corresponding to the acquisition target information.

The information acquisition unit 215 uses the position information corresponding to the frame from which the first information is acquired as a reference, and scans the other frame in a predetermined direction from the reference based on the position information of the other frame. The second information is acquired by searching.
This makes it possible to perform image analysis with higher accuracy, for example, in a table in which information related to the vertical direction or the horizontal direction is arranged.

The frame detection unit 213 separates and detects the frame line in the first direction and the frame line in the second direction, which are the frame lines of a plurality of frames, from the area corresponding to the table, and detects the detected first direction. A plurality of frames are detected based on the intersection of the borders of the above and the borders of the second direction.
In this way, by detecting the frame line in the first direction and the frame line in the second direction separately, the frame line and the frame are detected more accurately than in the case where the detection is performed without separating them. be able to.

The frame detection unit 213 cuts out each of the detected plurality of frames.
The character recognition result is generated by an image analysis process in which each of the plurality of cut out frames is used as a unit.
As a result, the image analysis process can be performed in units of the frame including the character string, so that the character recognition can be performed with higher accuracy.

The image is an image generated by taking one of a plurality of tables having different appearances in which the results of the medical examination are described.
The information to be acquired includes at least the results of health examinations.
As a result, the health diagnosis results described in various formats can be obtained with high accuracy.

[Realization of functions by hardware and software]
The function of executing a series of processes according to the above-described embodiment can be realized by hardware, software, or a combination thereof. In other words, it suffices if the function of executing the above-mentioned series of processes is realized in any one of the image analysis systems S, and the mode in which this function is realized is not particularly limited.

For example, when the function of executing the above-mentioned series of processes is realized by a processor that executes the arithmetic processing, the processor that executes the arithmetic processing is composed of various processing units such as a single processor, a multiprocessor, and a multicore processor. In addition to the above, the present invention includes a combination of these various processing units and a processing circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field-Programmable Gate Array).

Further, for example, when the function of executing the above-mentioned series of processes is realized by software, the programs constituting the software are installed in the computer via a network or a recording medium. In this case, the computer may be a computer having dedicated hardware built-in, or a general-purpose computer capable of performing a predetermined function by installing a program (for example, a general-purpose personal computer or the like). It may be an electronic device in general). Further, the step for describing the program may include only the processes performed in time series in the order thereof, but may include the processes executed in parallel or individually. Further, the steps for describing the program may be executed in any order within a range that does not deviate from the gist of the present invention.

The recording medium on which such a program is recorded may be provided to the user by being distributed separately from the computer main body, or may be provided to the user in a state of being incorporated in the computer main body in advance. In this case, the storage medium distributed separately from the computer itself is composed of, for example, a magnetic disk (including a floppy disk), an optical disk, a magneto-optical disk, or the like. The optical disc is composed of, for example, a CD-ROM (Compact Disc-Read Only Memory), a DVD (Digital Versaille Disc), a Blu-ray (registered trademark) Disc (Blu-ray disc), or the like. The magneto-optical disk is composed of, for example, MD (MiniDisc) or the like. These storage media are attached to, for example, the drive 15 of FIG. 2 and the drive 25 of FIG. 3 and incorporated into the computer body. The recording media provided to the user in a state of being incorporated in the computer body in advance include, for example, the ROM 12 of FIG. 2 and the ROM 22 of FIG. 3 in which the program is recorded, and the storage unit 16 and FIG. 3 of FIG. It is composed of an HDD, an SSD, or the like included in the storage unit 26.

1 terminal, 2 image analyzer, 3 manuscript (health diagnosis table), 11,21 CPU, 12,22 ROM, 13,23 RAM, 14,24 communication unit, 15,25 drive, 16,26 storage unit, 17, 27 Input unit, 18,28 output unit, 111 shooting control unit, 112 terminal side notification unit, 161 and 261 image storage unit, 162, 263 analysis result storage unit, 211 image acquisition unit, 212 table detection unit, 213 frame detection unit , 214 character recognition unit, 215 information acquisition unit, 216 device side notification unit, 262 dictionary data storage unit, S image analysis system

Claims (8)

A table detecting means for detecting an area corresponding to the table from an image including the table as a subject,
A frame detecting means for detecting a plurality of frames constituting the table from a region corresponding to the table and assigning position information in the image to each of the detected frames.
An information acquisition means for acquiring acquisition target information based on the character recognition result for each of the plurality of frames and the position information given to each of the plurality of frames.
An image analysis system characterized by being equipped with. The information acquisition means is
The first information is acquired based on the dictionary data about the first information which is a part of the acquisition target information and the character recognition result of each of the plurality of frames.
The second information, which is a part of the acquisition target information, is acquired based on the position information corresponding to the frame from which the first information is acquired.
The image analysis system according to claim 1. The information acquisition means uses the position information corresponding to the frame from which the first information is acquired as a reference, and scans another frame in a predetermined direction from the reference based on the position information of the other frame. The second information is acquired by searching for.
The image analysis system according to claim 2. The frame detecting means separates and detects the frame line in the first direction and the frame line in the second direction, which are the frame lines of the plurality of frames, from the region corresponding to the table, and detects the first frame line. The plurality of frames are detected based on the intersection of the borders in the direction of 1 and the borders in the second direction.
The image analysis system according to any one of claims 1 to 3, wherein the image analysis system is characterized by the above. The frame detecting means cuts out each of the plurality of detected frames,
The character recognition result is generated by an image analysis process in which each of the plurality of cut out frames is a unit.
The image analysis system according to any one of claims 1 to 4, wherein the image analysis system is characterized by the above. The image is an image generated by taking one of a plurality of tables having different appearances in which the results of the medical examination are described.
The acquisition target information includes at least the result of a health diagnosis.
The image analysis system according to any one of claims 1 to 5, wherein the image analysis system is characterized by the above. A table detection step for detecting an area corresponding to the table from an image including the table as a subject,
A frame detection step of detecting a plurality of frames constituting the table from the area corresponding to the table and assigning position information in the image to each of the detected frames.
An information acquisition step for acquiring acquisition target information based on the character recognition result for each of the plurality of frames and the position information given to each of the plurality of frames.
An image analysis method characterized by including. A table detection function that detects the area corresponding to the table from an image that includes the table as a subject,
A frame detection function that detects a plurality of frames constituting the table from the area corresponding to the table and assigns position information in the image to each of the detected frames.
An information acquisition function for acquiring acquisition target information based on the character recognition result for each of the plurality of frames and the position information given to each of the plurality of frames.
An image analysis program characterized by realizing the above on a computer.

Images