Inspection support system, inspection support method, and recording medium on which inspection support program is recorded

Abstract

An examination assistance system for assisting an examination using a medical image, the examination assistance system comprising one or more processors having hardware, the processors acquiring a video signal including information of the medical image, and identifying the medical image by performing an identification process based on the video signal. And a determination unit configured to output inspection assistance information for assisting the inspection, and determine whether or not the video signal is a video signal for the recognition processing on the basis of the video signal.

Description

Technical Field

[0001] The present invention relates to an inspection support system, an inspection support method and a recording medium recording an inspection support program suitable for equipment that performs AI determination. Background Art

[0002] In recent years, medical systems that use AI (artificial intelligence) to make various judgments have also been developed in the medical field. For example, in colon endoscopy, in order to perform high-quality endoscopy with less burden on patients, an endoscope AI operation assistance system is sometimes used. This system uses AI to judge the current endoscopy status and prompts auxiliary information that matches the status.

[0003] In order to perform high-precision AI judgment, it is necessary to use an endoscopic image suitable for AI judgment. However, even if an endoscopic image suitable for AI judgment is captured, if there is a wiring error between the video processing device of the endoscope system and the device that outputs operation assistance information based on the endoscopic image through AI (hereinafter referred to as the inspection assistance device), it is impossible to perform AI judgment with a suitable endoscopic image.

[0004] As a technology for preventing such wiring errors, Japanese Patent Application Laid-Open No. 2010-268063 discloses a technology for authenticating a device to be connected in order to manage whether information communication between electronic devices is permitted. In this proposal, when an image forming device detects the connection of an external device, the external device ID of the external device is obtained, and when the pre-stored external device ID matches the obtained external device ID, the use of the external device is permitted.

[0005] By applying this technology to a video processing device and an inspection support device, it is possible to confirm whether the connection destination device of the communication connection is correct.

[0006] Prior art literature

[0007] Patent Literature

[0008] Patent Document 1: Japanese Patent Application Publication No. 2010-268063 Summary of the invention

[0009] Problem that the invention aims to solve

[0010] However, between the video processing device and the inspection auxiliary device, two types of connections are used: a communication connection for transmitting an identification signal or a control signal, and an image connection for transmitting an image signal of an endoscopic image. Therefore, even if it is assumed that the technology of Patent Publication No. 2010-268063 is applied in the connection between the video processing device and the inspection auxiliary device to confirm that the connection destination device of the communication connection is correct, this cannot guarantee that the connection destination device of the image signal is correct. In addition, sometimes it is configured so that video output from multiple output ports can be output from the video processing device in each transmission format of the image. In this case, for example, even if the connection destination device of the image signal is correct, sometimes an image signal that is not used for AI judgment may be sent to the inspection auxiliary device due to incorrect wiring.

[0011] An object of the present invention is to provide an inspection support system, an inspection support method, and a recording medium recording an inspection support program, which are capable of checking whether an image signal input to an inspection support device is an appropriate image signal.

[0012] Solutions for solving problems

[0013] An inspection assistance system of one embodiment of the present invention is an inspection assistance system for assisting inspections using medical images, wherein the system has one or more processors with hardware, the processors obtaining an image signal containing information of the medical image, and outputting inspection assistance information for assisting the inspection by performing recognition processing based on the image signal, and determining whether the image signal is an image signal for the recognition processing based on the image signal.

[0014] An inspection assisting method according to one embodiment of the present invention is an inspection assisting method for assisting inspections using medical images, and comprises: obtaining an image signal containing information of the medical image; determining, based on the image signal, whether the image signal is an image signal for identification processing; and outputting inspection assisting information for assisting inspections using the medical image by performing identification processing based on the image signal.

[0015] A recording medium of one embodiment of the present invention records the examination assistance program, and the examination assistance program enables a computer to perform the following processing: processing of obtaining an image signal containing information of the medical image; processing of determining whether the image signal is an image signal for identification processing based on the image signal; and processing of outputting examination assistance information for assisting an examination using the medical image by performing identification processing based on the image signal.

[0016] Effects of the Invention

[0017] According to the present invention, there is an effect that it is possible to check whether the image signal input to the inspection supporting device is an appropriate image signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a diagram showing the configuration of an inspection support system according to a first embodiment of the present invention.

[0019] Figure 2 This is an explanatory diagram for explaining an example of a determination method of input eligibility determination performed by the AI ​​information confirmation unit 13b.

[0020] Figure 3 This is an explanatory diagram for explaining an example of a determination method of input eligibility determination performed by the AI ​​information confirmation unit 13b.

[0021] Figure 4 2 is a flowchart showing AI determination control performed by the AI ​​information confirmation unit 13b.

[0022] Figure 5 2 is a flowchart showing the device legitimacy determination performed by the connection destination confirmation unit 13a.

[0023] Figure 6 2 is a flowchart showing the device legitimacy determination performed by the connection destination confirmation unit 13a.

[0024] Figure 7 is a block diagram showing a modified example. DETAILED DESCRIPTION

[0025] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[0026] (First embodiment)

[0027] Figure 1 1 is a block diagram showing a test support system according to a first embodiment of the present invention. This embodiment is an embodiment in which a medical image suitable for AI determination is input to a test support device.

[0028] This embodiment applies the inspection support system to endoscopic inspection, but is not limited to endoscopic inspection, and can be applied to various AI processing using AI judgment. For example, it can be applied to AI processing that outputs endoscopic operation guidance information based on endoscopic images, AI processing that outputs control information for automatic insertion of an endoscope based on endoscopic images, and AI processing that outputs identification information of lesions or organ sites based on endoscopic images.

[0029] Figure 1The configuration of the inspection support device 10 constituting the inspection support system according to the first embodiment of the present invention and the video processing device 20 that outputs information used for AI determination to the inspection support device is shown.

[0030] The video processing device 20 includes a device information output unit 21 and an input information output unit 22. The video processing device 20 may be connected to, for example, an endoscope (not shown) that is inserted into a subject and takes endoscopic images related to the body cavity, and endoscopic images such as endoscopic inspection images are supplied from the endoscope. The video processing device 20 can not only obtain endoscopic images, but also various information. For example, the video processing device 20 can obtain information on whether the video processing device 20 is connected to an endoscope and whether an inspection by the endoscope has been started. In addition, the video processing device 20 can also obtain information related to the switching of various observation modes such as NBI (narrow band observation) mode and RDI (infrared light observation) mode of the endoscope.

[0031] The device information output unit 21 transmits the source device information related to the device (source device) that is the source of the input information of the inspection supporting device 10, that is, the video processing device 20 that is the connection destination of the inspection supporting device 10, via the first communication cable C1 that constitutes the first communication path. In addition, the source device information transmitted by the device information output unit 21 may include, for example, information related to the action of the video processing device 20 (action information), information related to the operation of the video processing device 20 (operation information), and information related to the image format of the image signal output from the video processing device 20, in addition to the ID information that specifies the video processing device 20. The action information and operation information in the source device information include information on whether the video processing device 20 is connected to the endoscope, information on whether the inspection by the endoscope has started, information related to the switching of the observation mode of the endoscope, and the like.

[0032] The input information output unit 22 outputs the input information of the inspection assisting device 10 via the second communication cable C2 constituting the second communication path. That is, the input information output unit 22 outputs an image for AI determination, for example, an image signal of an image including an endoscopic image such as an endoscopic inspection image, as input information of the inspection assisting device 10. In addition, the image based on the image signal output by the input information output unit 22 includes an endoscopic image as a medical image in a part of the image.

[0033] Furthermore, although it has been described that the first and second communication paths are respectively constituted by the first communication cable C1 and the second communication cable C2, the present invention is not limited thereto, and the first and second communication paths may be constituted by wireless paths.

[0034] The inspection auxiliary device 10 includes a device information acquisition unit 11, an input information acquisition unit 12, an input information confirmation unit 13, and an AI determination unit 14. The entire inspection auxiliary device 10, or at least one of the device information acquisition unit 11, the input information acquisition unit 12, the input information confirmation unit 13, and the AI ​​determination unit 14, may be composed of one or more processors using a CPU (Central Processing Unit) or an FPGA (Field Programmable Gate Array), or the like, or each unit may be controlled by operating in accordance with a program stored in a memory not shown, or a part or all of the functions may be realized by a hardware electronic circuit.

[0035] The device information acquisition unit 11 acquires the transmission source device information transmitted from the device information output unit 21 via the first communication path, and outputs the acquired transmission source device information to the input information confirmation unit 13. The input information acquisition unit 12 acquires the input information transmitted from the input information output unit 22 via the second communication path, and outputs the acquired input information to the input information confirmation unit 13.

[0036] The input information confirmation unit 13 includes a connection destination confirmation unit 13a and an AI information confirmation unit 13b. The connection destination confirmation unit 13a is provided with the transmission source device information from the device information output unit 21. The connection destination confirmation unit 13a uses the transmission source device information to determine the device legitimacy of the connection destination device, i.e., the device that is the transmission source of the endoscopic image as the input information for determining the content of assistance in the inspection assistance device 10, and confirms the legitimacy of the device.

[0037] The AI ​​information confirming unit 13b receives input information from the input information outputting unit 22. The AI ​​information confirming unit 13b performs an input eligibility determination on whether the input information is suitable for a predetermined identification process such as AI determination, and outputs the input eligibility determination result to the AI ​​determining unit 14.

[0038] The AI ​​determination unit 14 is provided with input information and an input eligibility determination result from the AI ​​information confirmation unit 13b. Based on the input eligibility determination result, the AI ​​determination unit 14 decides whether to use the input information for AI determination. That is, when the AI ​​determination unit 14 is provided with an input eligibility determination result that the input information is suitable for identification processing from the AI ​​information confirmation unit 13b, the AI ​​determination unit 14 performs identification processing such as AI determination based on the input information. The AI ​​determination unit 14 outputs the determination result of the AI ​​determination as the identification result of the identification processing to the connection destination confirmation unit 13a. In addition, when the AI ​​determination unit 14 is provided with an input eligibility determination result that the input information is not suitable for identification processing from the AI ​​information confirmation unit 13b, the AI ​​determination unit 14 does not perform identification processing such as AI determination based on the input information.

[0039] Furthermore, the AI ​​information confirmation unit 13b may supply the input information to the AI ​​determination unit 14 only when an input eligibility determination result indicating that the input information is suitable for recognition processing is provided.

[0040] When the connection destination confirmation unit 13a receives the AI ​​determination result from the AI ​​determination unit 14, it determines whether to adopt the AI ​​determination result based on the determination result of the device legitimacy of the source device of the input information used in the AI ​​determination. That is, when the connection destination confirmation unit 13a determines that the device is legitimate based on the determination result of the device legitimacy, it adopts the AI ​​determination result and outputs the AI ​​determination result to the display unit 30. On the other hand, when the connection destination confirmation unit 13a determines that the device is not legitimate based on the determination result of the device legitimacy, it does not output the AI ​​determination result to the display unit 30.

[0041] The display unit 30 displays the information from the input information confirmation unit 13 on the display screen. When the AI ​​determination result is input, the display unit 30 displays the input AI determination result on the display screen. The input information confirmation unit 13 may also provide the determination result of the legitimacy of the device and the determination result of the input eligibility to the display unit 30 for display.

[0042] (Input eligibility determination)

[0043] Figure 2 and Figure 3 This is an explanatory diagram for explaining an example of a determination method of input eligibility determination performed by the AI ​​information confirmation unit 13b.

[0044] The video processing device 20 generates an image signal of a predetermined image format in which image parameters such as resolution, color gamut, and frame rate are appropriately set by processing the predetermined image signal of the video signal obtained by shooting with an endoscope (not shown). The video processing device 20 encodes the generated image signal using a codec specified by MPEG-4 or H.265 / HEVC, etc., and containerizes the encoded image signal in various file formats such as AVI, MOV, WMV, and FLV, and transmits it via an output terminal that complies with a transmission standard specified by DVI or SDI, etc. In addition, the image format of the image signal that can be transmitted is predetermined in accordance with the transmission standard.

[0045] When an image format that is not suitable for AI determination is used, a highly reliable AI determination result cannot be obtained. In addition, even when an image format suitable for AI determination is input to the video processing device 20, an image that is not suitable for AI determination may be transmitted to the inspection support device 10 due to wiring errors or the like. In this case, a highly reliable AI determination result cannot be obtained.

[0046] Figure 2 The image P1 shows an example of an image suitable for AI determination by the AI ​​determination unit 14, which is obtained by imaging the image signal output from the input information output unit 22. The image P1 has a display area P1R for displaying the endoscopic image in a relatively large range of the entire image area. The display area P1R includes an endoscopic image P1Ra in which the shape of the effective image area is an octagonal shape by a masking area (full coating part) whose four corners are triangles. The endoscopic image P1Ra has a relatively high resolution. The AI ​​determination unit 14 obtains an AI determination result by performing recognition processing on the endoscopic image P1Ra.

[0047] In contrast, Figure 2The image P2 shows an example of an image that is not suitable for AI determination by the AI ​​determination unit 14. For example, when the type of endoscope connected to the video processing device 20 is inappropriate or the setting of the video processing device 20 is incorrect, when the image signal from the input information output unit 22 is imaged, it is conceivable that there will be a situation where the image P2 is obtained. That is, the image P2 includes an endoscopic image P1Rb in which the shape of the effective image area is an octagonal shape by a masking area (full coating part) with four triangular corners in a display area P1R of a relatively large range in the entire image area. The aspect ratio of the endoscopic image P1Rb is different from that of the endoscopic image P1Ra. That is, the positions of the endoscopic image P1Ra and the endoscopic image P1Rb in the image based on the image signal as the input information are different from each other, and even if the AI ​​determination unit 14 performs recognition processing on the endoscopic image P1Rb, it may not be able to obtain a correct AI determination result.

[0048] In addition, in the case where the endoscope to be connected to the video processing device 20 is connected by mistake or the setting is incorrect, or in the case where an image not for AI determination, such as an image for a printer, is input to the inspection auxiliary device 10, the resolution of the endoscopic image P1Rb input to the inspection auxiliary device 10 may be lower than the resolution of the endoscopic image P1Ra. In this case, even if the AI ​​determination unit 14 performs recognition processing on the endoscopic image P1Rb, it may not necessarily obtain a correct AI determination result. Moreover, in the case where an image to be output to another device not shown is mistakenly provided to the inspection auxiliary device 10, another image P2a may be superimposed on a part of the endoscopic image P1Rb through PIP (picture-in-picture). In this case, the AI ​​determination unit 14 may not obtain a correct AI determination result.

[0049] The AI ​​information confirmation unit 13b can also determine that the image signal is not the image signal for the identification processing in at least one of the following situations: the aspect ratio or image quality of the learning medical image used in generating the inference model of the AI ​​judgment unit 14 is different from at least one of the aspect ratio and image quality of the medical image serving as input information; and in at least one of the following situations: the medical image is superimposed with other images.

[0050] In order to make such a determination, for example, the AI ​​information confirmation unit 13b may also have a memory (not shown) that stores information related to the features of an image suitable for AI determination. The AI ​​information confirmation unit 13b compares the information in the memory with the input information by analyzing the image of the input information or the metadata attached to the input information, regarding the format of the image quality, frame rate, aspect ratio, superposition of other information, and other features, thereby determining the eligibility of the input information and obtaining an input eligibility determination result. For example, the AI ​​information confirmation unit 13b can determine the position of a medical image in an image based on an image signal as input information, and obtain an input eligibility determination result based on the consistency between the position and the position of the medical image in an image suitable for AI determination.

[0051] Figure 2 An example of determining the eligibility of input information by image analysis of an endoscopic image as input information has been described. Figure 3 An example is shown in which the input information includes information directly indicating the eligibility of the input information.

[0052] exist Figure 3 In the image P3, identification information P3a indicating that the endoscopic image P1Ra is an image suitable for AI determination is superimposed in the display area P1R. Figure 3 In the example of , the identification information P3a is image information. The identification information P3a is arranged in the masked areas at the four corners of the display area P1R, and the identification information P3a hardly causes information loss in the effective image area of ​​the endoscopic image P1Ra.

[0053] As the identification information P3a, for example, a barcode or a QR code (registered trademark, Quick Response Code) can be used. Figure 3 In FIG. 1 , an example in which the identification information P3a can be visually recognized is shown, but identification information that cannot be visually recognized may be used as the identification information P3a.

[0054] The identification information P3a may be included not only in an image based on an image signal but also in header information of an image signal, or metadata indicating that the endoscopic image P1Ra is an image suitable for AI determination may be included in the input information.

[0055] The AI ​​information confirmation unit 13b can determine the validity of the input information by confirming the identification information P3a. In this way, the AI ​​information confirmation unit 13b determines whether the image signal is suitable for recognition processing such as AI determination based on the image signal.

[0056] In addition, the AI ​​information confirmation unit 13b performs input eligibility determination at an appropriate timing. For example, the AI ​​information confirmation unit 13b may perform input eligibility determination when the inspection auxiliary device 10 is started, or when an endoscope (not shown) is connected to the video processing device 20. In addition, the AI ​​information confirmation unit 13b may perform input eligibility determination when the setting of the video processing device 20 that supplies input information to the inspection auxiliary device 10 is changed, or may perform input eligibility determination at a predetermined period.

[0057] When the AI ​​information confirmation unit 13b determines that the input information is not suitable for AI judgment based on the result of the input eligibility judgment, it can also generate display information for displaying a message (alarm) such as "This is not the target endoscope" or "The settings are different" on the display screen of the display unit 30 to remind people to pay attention, and supply it to the display unit 30.

[0058] (Equipment Legitimacy Determination)

[0059] In order to determine the legitimacy of the device, the connection destination confirmation unit 13a may also have a memory storing ID information related to a legitimate device. The connection destination confirmation unit 13a may also determine the legitimacy of the device by comparing the information stored in the memory with the source device information. For example, in the case where endoscopic images are obtained in multiple examination rooms, it is conceivable that an endoscopic image obtained in an examination room different from the target examination room is input to the AI ​​information confirmation unit 13b from the video processing device 20 as input information due to wiring errors or the like. Since the AI ​​judgment for the input image can be correctly obtained in this case, if the source device information is not confirmed, operation assistance will be performed based on the endoscopic image of the examination room different from the target examination room. In this embodiment, the connection destination confirmation unit 13a compares the information stored in the memory with the source device information to determine the legitimacy of the device. In the case where the ID information stored in the memory is different from the ID information from the video processing device 20, the AI ​​information confirmation unit 13b determines that the source device is not legitimate, thereby preventing the user from receiving erroneous operation assistance.

[0060] In addition, as described above, the source device information received by the connection destination confirmation unit 13a includes not only the ID information of the video processing device 20, but also the information of the endoscope to be connected, the action information and operation information of the video processing device 20 including the information related to the examination, and the information related to the image format. The connection destination confirmation unit 13a can also determine the legitimacy of the device based on the information such as the action information, operation information and image format. Furthermore, the connection destination confirmation unit 13a can also determine the legitimacy of the device using the information from the AI ​​information confirmation unit 13b or the result of the AI ​​determination.

[0061] For example, when the endoscopic image of the input information should change due to a change in the operation of the video processing device 20, the connection destination confirmation unit 13a may determine that the device is not legitimate when the AI ​​information confirmation unit 13b determines that the same image is continuously input, and may determine that the device is legitimate when the AI ​​information confirmation unit 13b determines that the image has changed. Alternatively, the connection destination confirmation unit 13a may determine that the device is legitimate when the information of the image format is consistent with the information of the image format determined by the AI ​​information confirmation unit 13b, and may determine that the device is not legitimate when the two are inconsistent. In this way, the connection destination confirmation unit 13a can determine the legitimacy of the device based on the source device information from the video processing device 20 and the image recognition result of the AI ​​information confirmation unit 13b.

[0062] In addition, the connection destination confirmation unit 13a may determine that the device is legitimate when, for example, operation information indicating the start of endoscopic examination is input, and further, through the AI ​​judgment of the AI ​​judgment unit 14, a recognition result is obtained that the input image has changed to an image in which the body is photographed, that is, the examination has started. In addition, the connection destination confirmation unit 13a may determine the legitimacy of the device based on whether, for example, operation information related to an observation mode such as NBI or RDI is consistent with the observation mode obtained by the AI ​​judgment result.

[0063] That is, the video processing device 20 as a medical image generating device can set an observation mode for the endoscope. Based on the AI ​​judgment result, it is determined whether the endoscopic image obtained in the observation mode set by the video processing device 20 has been supplied as input information to the inspection auxiliary device 10. That is, the legitimacy of the device is determined by determining whether the medical image generating device at the connection destination of the first communication path and the medical image generating device at the connection destination of the second communication path are the same medical image generating device.

[0064] Furthermore, when the connection destination confirmation unit 13 a determines that the device is not authorized through the device authorization determination, the connection destination confirmation unit 13 a may output a signal for stopping the operation of the AI ​​determination unit 14 to the AI ​​determination unit 14 .

[0065] (Various judgments)

[0066] Alternatively, the AI ​​determination unit 14 can perform various AI determinations. For example, the AI ​​determination unit 14 can output operation guidance information of the endoscope as inspection auxiliary information. Alternatively, the AI ​​determination unit 14 can output control information for automatically inserting the endoscope. Alternatively, the AI ​​determination unit 14 can output identification information of the lesion or organ site.

[0067] The input information confirmation unit 13 may also control the AI ​​determination unit 14 in a manner that does not output the inspection auxiliary information when the device legitimacy judgment indicating that the source device is not legitimate or the input qualification judgment indicating that the input information is input information that is not suitable for AI determination is obtained. In addition, when the endoscope connected to the video processing device 20 is an automatically controlled endoscope that can be automatically inserted, the input information confirmation unit 13 may be configured to operate an emergency stop device (not shown) for stopping the automatic insertion of the automatically controlled endoscope when the device legitimacy judgment indicating that the source device is not legitimate or the input qualification judgment indicating that the input information is input information that is not suitable for AI determination is obtained.

[0068] That is, when the input information confirmation unit 13 determines that the source device of the connection destination of the first communication path and the source device of the connection destination of the second communication path are the same source device, the inspection auxiliary information is output. In addition, when the input information confirmation unit 13 determines that the source device of the connection destination of the first communication path and the source device of the connection destination of the second communication path are different, the alarm information related to the connection destination of the first communication path and the connection destination of the second communication path is output.

[0069] (action)

[0070] Next, refer to Figures 4 to 6 The operation of the embodiment configured in this way will be described. Figure 4 2 is a flowchart showing AI determination control performed by the AI ​​information confirmation unit 13b. Figure 5 and Figure 6 2 is a flowchart showing the device legitimacy determination performed by the connection destination confirmation unit 13a.

[0071] The endoscope is connected to the video processing device 20, and the video processing device 20 drives the endoscope to obtain an endoscopic image. For example, in order to assist in the operation of endoscopic inspection, the input information output unit 22 of the video processing device 20 outputs the endoscopic image obtained by the endoscope to the inspection assisting device 10 via the second communication path. In addition, the device information output unit 21 of the video processing device 20 outputs the transmission source device information including the action information, operation information of the video processing device 20, and the image format of the endoscopic image to the inspection assisting device 10 via the first communication path.

[0072] exist Figure 4 In S1, the AI ​​information confirmation unit 13b receives input information from the input information output unit 22 of the video processing device 20 in order to determine the input eligibility of the input information. The AI ​​information confirmation unit 13b determines whether the input information is an image suitable for AI determination by performing image analysis on the medical image or analyzing the metadata of the input information as the received input information (S2). When the AI ​​information confirmation unit 13b determines that the input information is an image suitable for AI determination (the determination of S2 is "yes"), the input eligibility determination result for causing the AI ​​determination unit 14 to implement AI determination is output to the AI ​​determination unit 14 (S3). In addition, when the AI ​​information confirmation unit 13b determines that the input information is not an image suitable for AI determination (the determination of S2 is "no"), the input eligibility determination result for not causing the AI ​​determination unit 14 to implement AI determination is output to the AI ​​determination unit 14 (S4).

[0073] In this case, the AI ​​determination unit 14 does not perform AI determination. In addition, the input information confirmation unit 13 displays a message to call attention, such as “This is not the target endoscope” or “The settings are different”, on the display screen of the display unit 30 .

[0074] When the AI ​​information confirmation unit 13b determines that the input information is suitable for AI determination, it provides the input information to the AI ​​determination unit 14. The AI ​​determination unit 14 obtains an AI determination result by performing recognition processing on the input information. The AI ​​determination unit 14 outputs the AI ​​determination result to the input information confirmation unit 13.

[0075] exist Figure 5In S11 of the input information confirmation unit 13, the connection destination confirmation unit 13a of the input information confirmation unit 13 receives the source device information from the device information output unit 21 of the video processing device 20 in order to determine the legitimacy of the device of the source device. The connection destination confirmation unit 13a determines whether the source device that sent the source device information is the target device by comparing the received source device information with the information stored in the memory (S12). When the connection destination confirmation unit 13a determines that the source device is legitimate (the determination of S12 is "yes"), the AI ​​determination result from the AI ​​determination unit 14 is displayed on the display unit 30 (S13). In addition, when the connection destination confirmation unit 13a determines that the source device is not legitimate (the determination of S12 is "no"), the AI ​​determination result from the AI ​​determination unit 14 is not displayed on the display unit 30 (S14).

[0076] Figure 5 The example shows an example in which the connection destination confirmation unit 13a determines the legitimacy of the sending source device based on the ID information of the sending source device and decides whether to display or not display the AI ​​judgment result, but the connection destination confirmation unit 13a can also use the output of the AI ​​information confirmation unit 13b and the AI ​​judgment unit 14 to determine the legitimacy of the sending source device and decide whether to display or not display the AI ​​judgment result.

[0077] Figure 6 An example of this case is shown. Figure 6 In S31 of the input information confirmation unit 13, the AI ​​information confirmation unit 13b of the input information confirmation unit 13 determines whether the input information is suitable for AI determination. When the input information is not suitable for AI determination, as described above, AI determination is not performed. When the input information is suitable for AI determination, AI determination is performed and the AI ​​determination result is provided to the input information confirmation unit 13.

[0078] In S32, the input information confirmation unit 13 determines the legitimacy of the source device based on the ID information and the like (source determination). When the input information confirmation unit 13 determines that the source device is legitimate (S33 is determined as "yes"), it determines whether the inspection has started in the next S34. That is, when the input information confirmation unit 13 detects that the inspection has started based on the source device information, it obtains the AI ​​determination result of whether the inspection has started from the AI ​​determination unit 14. When both the source device information and the AI ​​determination result indicate that the inspection has started (S35 is determined as "yes"), the input information confirmation unit 13 determines the observation mode in the next S36.

[0079] When the input information confirmation unit 13 detects the observation mode based on the transmission source device information, it obtains the AI ​​determination result about the observation mode from the AI ​​determination unit 14. When the transmission source device information and the AI ​​determination result both indicate the same observation mode (the determination of S37 is "yes"), the input information confirmation unit 13 determines that the transmission source is a legitimate device in the next S38. In this case, the input information confirmation unit 13 displays the AI ​​determination result from the AI ​​determination unit 14 on the display unit 30.

[0080] On the other hand, when the input information confirmation unit 13 determines "No" in S33, S35, and S37, the input information confirmation unit 13 determines in S39 that the transmission source is not a legitimate device. In this case, the input information confirmation unit 13 does not display the AI ​​determination result from the AI ​​determination unit 14 on the display unit 30. In addition, in this case, the input information confirmation unit 13 displays a message such as "It is not the target endoscope" and "The settings are different" on the display screen of the display unit 30 to warn.

[0081] Thus, in this embodiment, since not only the legitimacy of the source device but also the eligibility of the input information is determined, it is possible to reliably determine that the input information suitable for AI determination is input from the target device. This prevents the display of inappropriate AI determination results in the case where an inappropriate image is input during AI processing due to a connection error or the like, or in the case where AI determination is performed on input information from a device that is not the target.

[0082] (Variation Example)

[0083] Figure 7 FIG. 1 is a block diagram showing a modified example. Figure 1 In the example in which the video processing device 20 transmits the transmission source device information and the input information to the inspection assisting device 10, it is described. Figure 7 In a modified example, when identification information is superimposed on input information and transmitted, information similar to the identification information is transmitted to the inspection supporting device 10 through the third communication cable C3 constituting the third communication path.

[0084] The video processing device 20A is a video processing device to which an identification information output unit 23 is added. When identification information is added to input information from the device information output unit 21 and sent, the identification information output unit 23 outputs identification information having the same content as the identification information added to the input information to the inspection supporting device 10 via the third communication cable C3. The device information acquisition unit 11 of the inspection supporting device 10 receives the identification information output from the identification information output unit 23 and provides it to the input information confirmation unit 13.

[0085] Furthermore, although the example in which the identification information output unit 23 outputs the identification information to the inspection supporting device 10 via the third communication cable C3 is shown, the third communication cable C3 may be omitted and the identification information may be supplied to the inspection supporting device 10 via the first communication cable C1 .

[0086] The device information acquisition unit 11 of the inspection auxiliary device 10 provides the received identification information to the AI ​​information confirmation unit 13b. The AI ​​information confirmation unit 13b compares the identification information added to the input information with the identification information received via the third communication cable C3. When it is determined that these information are information of the same content, it can be determined that the transmission source of the input information is consistent with the transmission source of the transmission source device. Depending on the situation, it can also be further determined that the input information is information suitable for AI determination.

[0087] Other components and functions are related to Figure 1 The implementation method is the same.

[0088] According to the modified example configured in this way, since identification information of the same content is transmitted via the first communication path and the second (or third) communication path, the legitimacy of the transmission source of the input information can be determined relatively easily.

[0089] The present invention is not limited to the above-mentioned embodiments, and can be embodied by deforming the constituent elements within the scope of its main purpose in the implementation stage. In addition, various inventions can be formed by appropriate combinations of multiple constituent elements disclosed in the above-mentioned embodiments. For example, some constituent elements of all constituent elements shown in the embodiments can also be deleted. Moreover, constituent elements in different embodiments can also be appropriately combined.

Claims

1. An inspection support system is an inspection support system for supporting inspections using medical images, characterized in that: Have more than one processor with hardware, The processor obtaining an image signal containing information of the medical image, By performing recognition processing based on the image signal, inspection auxiliary information for assisting the inspection can be output. Based on the video signal, it is determined whether the video signal is the video signal for the recognition process.

2. The inspection assistance system according to claim 1, characterized in that: The processor Acquiring source device information related to the source device via a first communication path connected to the source device, and determining whether the source device at the connection destination is appropriate based on the source device information, wherein the source device is a source of the medical image obtained by photographing the subject. The video signal is acquired via a second communication path connected to the transmission source device, and based on the video signal, it is determined whether the video signal is the video signal for the recognition process.

3. The inspection assistance system according to claim 1, characterized in that: The processor determines whether the image signal is the image signal for the identification processing based on the position of the medical image on the image obtained by imaging the image signal.

4. The inspection assistance system according to claim 1, characterized in that: The processor determines whether the image signal is the image signal for the recognition processing based on the aspect ratio of the medical image.

5. The inspection assistance system according to claim 1, characterized in that: The processor determines whether the image signal is the image signal for the identification processing based on the image quality of the medical image.

6. The inspection assistance system according to claim 1, characterized in that: The processor determines whether the image signal is the image signal for the identification processing based on whether an image different from the medical image is superimposed on the medical image when the image signal is imaged.

7. The inspection assistance system according to claim 1, characterized in that: The processor obtaining identification information contained in the image signal that is different from the information of the medical image, Based on the identification information, it is determined whether the video signal is the video signal for the recognition process.

8. The inspection assistance system according to claim 7, characterized in that: The identification information is a QR code.

9. The inspection assistance system according to claim 7, characterized in that: The processor obtaining information related to the content of the identification information via a first communication path connected to a source device that is a source of transmission of the medical image, acquiring the identification information via a second communication path connected to the transmission source device, Based on the information related to the content of the identification information obtained via the first communication path and the identification information obtained via the second communication path, it is determined whether the source device of the connection destination of the first communication path and the source device of the connection destination of the second communication path are the same source device.

10. The inspection assistance system according to claim 2, characterized in that: The sending source device is a medical image generating device that generates the medical image by setting an observation mode and observing the subject in the set observation mode. The processor obtaining information related to setting of the observation mode via the first communication path, acquiring the image signal via the second communication path, Based on whether the image signal corresponds to the setting of the observation mode, it is determined whether the medical image generation device at the connection destination of the first communication path and the medical image generation device at the connection destination of the second communication path are the same medical image generation device.

11. The inspection assistance system according to claim 1, characterized in that: The processor outputs an alarm when determining that the image signal is not the image signal for the recognition process.

12. The inspection assistance system according to claim 1, characterized in that: The processor When it is determined that the image signal is not an image signal for the recognition process, the inspection auxiliary information is not outputted. When it is determined that the image signal is the image signal for the recognition process, the inspection auxiliary information is output.

13. The inspection assistance system according to claim 9, characterized in that: The processor When it is determined that the transmission source device of the connection destination of the first communication path and the transmission source device of the connection destination of the second communication path are the same transmission source device, the inspection auxiliary information is outputted; When it is determined that the transmission source device of the connection destination of the first communication path is different from the transmission source device of the connection destination of the second communication path, alarm information about the connection destination of the first communication path and the connection destination of the second communication path is output.

14. The inspection assistance system according to claim 10, characterized in that: The processor When it is determined that the transmission source device of the connection destination of the first communication path and the transmission source device of the connection destination of the second communication path are the same transmission source device, the inspection auxiliary information is outputted; When it is determined that the transmission source device of the connection destination of the first communication path is different from the transmission source device of the connection destination of the second communication path, alarm information about the connection destination of the first communication path and the connection destination of the second communication path is output.

15. The inspection assistance system according to claim 1, characterized in that: The processor The recognition processing is performed by inputting the image signal into a recognition processing device that has been trained using a learning medical image, In at least one of the cases where the aspect ratio and image quality of the medical image are different from the aspect ratio and image quality of the learning medical image, and in at least one of the cases where other images are superimposed on the medical image, it is determined that the image signal is not the image signal for the identification processing.

16. The inspection assistance system according to claim 1, characterized in that: The processor outputs operation guidance information of the endoscope as the inspection auxiliary information.

17. The inspection assistance system according to claim 1, characterized in that: The processor outputs control information for automatically controlling the endoscope as the inspection auxiliary information.

18. An examination assistance method is a method for assisting an examination using a medical image, characterized in that: have: Obtaining an image signal containing information of the medical image; Based on the image signal, determining whether the image signal is the image signal for the recognition process; and By performing recognition processing based on the image signal, examination auxiliary information for assisting an examination using the medical image is output.

19. A recording medium having a program recorded thereon, the program causing a computer to execute the following processing: Obtaining processing of image signals containing information of the medical image; Based on the image signal, determining whether the image signal is an image signal for the recognition process; and A process of outputting examination auxiliary information for assisting an examination using the medical image by performing recognition processing based on the image signal.

Images