Blood storage level notification system, blood storage level notification device, and blood storage level notification method

Abstract

The present invention provides a blood storage level notification system, a blood storage level notification device, and a blood storage level notification method that can improve notification accuracy in accordance with hemodynamics. [Solution] A blood storage level notification system 100 for notifying the blood storage level of a blood storage reservoir tank 3, characterized in that it photographs the blood storage reservoir tank 3 to acquire image data, analyzes the image data to detect the blood storage level in the blood storage reservoir tank, predicts the future blood storage level based on the temporal change in the blood storage level, sets a timing for the blood storage reservoir tank 3 to reach a predetermined threshold, and provides notification corresponding to the prediction a predetermined time before the timing.

Description

Technical Field

[0005] ,

[0001] The present invention relates to a blood storage level notification system, a blood storage level notification device, and a blood storage level notification method for notifying the blood storage level of a blood storage reservoir tank.

Background Art

[0002] Conventionally, as a system for imaging the blood storage level of a blood storage reservoir tank with a USB network camera and extracting the blood storage level of the blood storage reservoir tank over time using the captured image data, for example, a system for extracting the blood level in a blood storage tank disclosed in Non-Patent Document 1 has been proposed.

[0003] The system for extracting the blood level in the blood storage tank disclosed in Non-Patent Document 1 is composed of a USB camera and a computer. An image of the blood storage tank is acquired from the USB camera connected to the computer, and image processing for calculating and displaying the change in the blood volume using binarization and the difference method is performed. The change in the liquid volume in the blood storage tank is calculated over time, and the change in the blood volume of the blood storage tank can be displayed.

Prior Art Documents

Non-Patent Documents

[0004]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] Here, regarding notification of the blood level in the blood reservoir tank, it is extremely important for the operator to maintain a constant blood level and blood flow rate in the reservoir tank during extracorporeal circulation. This can prevent problems with blood withdrawal and blood return. Therefore, it is necessary to maintain an appropriate perfusion flow rate and improve the accuracy of notification according to hemodynamics.

[0006] In this regard, Non-Patent Document 1 describes a method of acquiring blood reservoir levels using a USB camera, binarizing the image to black and white, and calculating and displaying the change in blood volume using the difference method. However, measures to prevent light interference are necessary when detecting blood reservoir levels. Furthermore, the display according to hemodynamics simply shows the change, and there is a concern that it will not be able to notify the operator in advance of the danger of the blood reservoir level, leading to a delayed response. Given these circumstances, there is a need to more accurately evaluate the blood reservoir level in the blood reservoir tank and improve the accuracy of notification according to hemodynamics.

[0007] Therefore, the present invention was devised in view of the above-mentioned problems, and its objective is to provide a blood storage level notification system, a blood storage level notification device, and a blood storage level notification method that can more accurately evaluate the blood storage level of a blood storage reservoir tank and improve the notification accuracy according to hemodynamics. [Means for solving the problem]

[0008] The blood storage level notification system according to the first invention is a blood storage level notification system for notifying the blood storage level of a blood storage reservoir tank, comprising: an image acquisition means for photographing the blood storage reservoir tank and acquiring image data; a blood storage level detection means for analyzing the image data and detecting the blood storage level in the blood storage reservoir tank; and a means for detecting the temporal change of the blood storage level. Based on the additionally entered patient biometric information and surgical progress data, The system is characterized by comprising: prediction means for predicting future blood storage levels; timing setting means for setting a timing at which the blood storage reservoir tank reaches a predetermined threshold based on the prediction; and notification means for providing notification corresponding to the prediction a predetermined time before the timing.

[0009] The blood storage level notification system according to the second invention is characterized in that, in the first invention, the blood storage level detection means analyzes the image data using an image recognition AI to detect the blood storage level in the blood storage reservoir tank, the prediction means compares the temporal change in the blood storage level with past blood storage level data using a machine learning model to predict the future blood storage level, and the timing setting means sets the timing at which the blood storage reservoir tank reaches a predetermined threshold based on the prediction by the machine learning model.

[0010] The blood storage level notification system according to the third invention, in the first invention, the blood storage level detection means based on the color change of the blood in the HSV color space and the set ROI of the image data. Numerical data showing the time-series fluctuations in the blood storage level in the blood storage reservoir tank. The system is characterized by detecting the blood storage level.

[0012] Fourth Invention The blood donation level notification system relating to the first invention is characterized in that the prediction means uses the skill information of the responder as additional input to improve the prediction accuracy.

[0013] Invention 5 The blood storage level notification system relating to the first invention is characterized in that the notification means provides notification information including the remaining time until the blood storage reservoir tank becomes empty or full, and information regarding the current hemodynamic status.

[0014] Invention #6 The blood storage level notification device is a blood storage level notification device that notifies the blood storage level of a reservoir tank, and comprises an image acquisition unit that photographs the blood storage reservoir tank and acquires image data, a blood storage level detection unit that analyzes the image data and detects the blood storage level in the blood storage reservoir tank, and the temporal change of the blood storage level Based on the additionally entered patient biometric information and surgical progress data, The system is characterized by comprising: a prediction unit that predicts future blood storage levels; a timing setting unit that sets the timing at which the blood storage reservoir tank reaches a predetermined threshold based on the prediction; and a notification unit that provides notification corresponding to the prediction a predetermined time before the timing.

[0015] Invention #7 The blood storage level notification device relating to this is Invention #6 In this system, the blood storage level detection unit analyzes the image data using image recognition AI to detect the blood storage level in the blood storage reservoir tank; the prediction unit compares the temporal change in the blood storage level with past blood storage level data using a machine learning model to predict the future blood storage level; and the timing setting unit sets the timing at which the blood storage reservoir tank reaches a predetermined threshold based on the prediction by the machine learning model.

[0016] Invention #8 The blood storage level notification method relating to the present invention is a blood storage level notification method for notifying the blood storage level of a blood storage reservoir tank, comprising: an image acquisition step of photographing the blood storage reservoir tank and acquiring image data; a blood storage level detection step of analyzing the image data and detecting the blood storage level in the blood storage reservoir tank; and the temporal change of the blood storage level. Based on the additionally entered patient biometric information and surgical progress data, The system is characterized by comprising: a prediction step of predicting future blood storage levels; a timing setting step of setting a timing at which the blood storage reservoir tank reaches a predetermined threshold based on the prediction; and a notification step of providing notification corresponding to the prediction a predetermined time before the timing.

[0017] 9th Invention The method for notifying blood storage level is as follows: Invention #8 In this system, the blood storage level detection step involves analyzing the image data using an image recognition AI to detect the blood storage level in the blood storage reservoir tank; the prediction step involves comparing the temporal change in the blood storage level with past blood storage level data using a machine learning model to predict the future blood storage level; and the timing setting step involves setting the timing at which the blood storage reservoir tank reaches a predetermined threshold based on the prediction by the machine learning model. [Effects of the Invention]

[0018] First Invention ~ Invention 5According to this, the blood storage level detection means analyzes the image data and detects the blood storage level in the blood storage reservoir tank. Therefore, the prediction means can predict the future blood storage level based on the temporal change of the blood storage level. As a result, the blood storage level of the blood storage reservoir tank can be evaluated more accurately, and the notification accuracy according to the blood dynamics can be improved.

[0019] In particular, according to the first invention, the timing setting means sets the timing when the blood storage reservoir tank reaches a predetermined threshold value based on the prediction. Therefore, the notification means can perform a notification corresponding to the prediction a predetermined time before the timing. As a result, the blood storage level of the blood storage reservoir tank can be evaluated more accurately, and the notification accuracy according to the blood dynamics can be improved. Furthermore, it becomes possible to notify the operator in advance of the corresponding actions to be noted, dangerous phenomena, etc.

[0020] In particular, according to the second invention, the blood storage level detection means can analyze the image data by an image recognition AI and detect the blood storage level in the blood storage reservoir tank. Therefore, the prediction means can collate the temporal change of the blood storage level and the past blood storage level data by a machine learning model and predict the future blood storage level. As a result, the blood storage level of the blood storage reservoir tank can be evaluated more accurately, and the notification accuracy according to the blood dynamics can be improved.

[0021] In particular, according to the third invention, the blood storage level detection means detects the blood storage level based on the color tone change in the HSV color space of the blood in the image data and the set ROI. Therefore, the prediction means can predict the future blood storage level based on the temporal change of the blood storage level. As a result, the blood storage level of the blood storage reservoir tank can be evaluated more accurately, and the notification accuracy according to the blood dynamics can be improved.

[0022] In particular, First InventionAccording to this, the prediction means uses the patient's biological information and the surgical progress status data as additional inputs. Therefore, the timing setting means can set the timing when the blood storage reservoir tank reaches a predetermined threshold value based on the prediction by the machine learning model. Thereby, the blood storage level of the blood storage reservoir tank can be evaluated more accurately, and the notification accuracy according to the blood dynamics can be improved.

[0023] In particular, Fourth Invention According to this, the prediction means uses the skill information of the operator as an additional input. Therefore, the timing setting means can set the timing when the blood storage reservoir tank reaches a predetermined threshold value based on the prediction by the machine learning model. Thereby, the blood storage level of the blood storage reservoir tank can be evaluated more accurately, and the notification accuracy according to the blood dynamics can be improved.

[0024] In particular, Fifth Invention According to this, the notification means includes the remaining time until the blood storage reservoir tank becomes empty or full and information regarding the current blood dynamics. Therefore, it is possible to notify the operator in advance of the corresponding actions that need attention, dangerous phenomena, etc. Thereby, the blood storage level of the blood storage reservoir tank can be evaluated more accurately, and the notification accuracy according to the blood dynamics can be improved.

[0025] Also, Invention #6 According to this, the blood storage level detection unit analyzes the image data and detects the blood storage level in the blood storage reservoir tank. Therefore, the prediction unit can predict the future blood storage level Based on the additionally entered patient biometric information and surgical progress data, from the temporal change of the blood storage level. Thereby, the blood storage level of the blood storage reservoir tank can be evaluated more accurately, and the notification accuracy according to the blood dynamics can be improved.

[0026] In particular, Invention 7According to the system, the blood reservoir level detection unit can detect the blood level in the reservoir tank by analyzing image data using image recognition AI. Therefore, the prediction unit can predict future blood levels by comparing the temporal changes in the blood reservoir level with past blood reservoir level data using a machine learning model. This allows for a more accurate assessment of the blood reservoir level in the reservoir tank and improves the accuracy of notifications in accordance with hemodynamic conditions.

[0027] Also, Invention #8 According to the documentation, the blood storage level detection step involves analyzing image data to detect the blood storage level in the blood storage reservoir tank. Therefore, the prediction step involves predicting the temporal change in the blood storage level. Based on the additionally entered patient biometric information and surgical progress data, It is possible to predict future blood storage levels. This allows for a more accurate assessment of blood storage levels in the blood storage reservoir tank, improving the accuracy of notifications based on hemodynamics.

[0028] especially, 9th Invention According to the system, the blood storage level detection step analyzes image data using image recognition AI to detect the blood storage level in the blood storage reservoir tank. Therefore, the prediction step compares the temporal changes in the blood storage level with past blood storage level data using a machine learning model to predict future blood storage levels. This allows for a more accurate assessment of the blood storage level in the blood storage reservoir tank and improves the accuracy of notifications in accordance with hemodynamics. [Brief explanation of the drawing]

[0029] [Figure 1] Figure 1 is a schematic diagram showing an example of a blood storage level notification system in this embodiment. [Figure 2] Figure 2(a) is a schematic diagram showing an example of the configuration of the blood storage level notification device in this embodiment, and Figure 2(b) is a schematic diagram showing an example of the function of the blood storage level notification device in this embodiment. [Figure 3] Figure 3 is a schematic diagram showing an example of the threshold timing in this embodiment. [Figure 4]Figure 4 is a schematic diagram showing an example of a reference database in this embodiment. [Figure 5] Figure 5 is a schematic diagram showing an example of a reference database in this embodiment. [Figure 6] Figure 6 is a flowchart showing an example of the operation of the blood storage level notification system in this embodiment. [Modes for carrying out the invention]

[0030] Hereinafter, an example of a blood storage level notification system and a blood storage level notification device in which the present invention is applied will be described with reference to the drawings.

[0031] (Embodiment: Blood storage level notification system 100) Referring to Figure 1, an example of the blood storage level notification system 100 in this embodiment will be described.

[0032] The blood level notification system 100 in this embodiment includes a blood level notification device 1, as shown in Figure 1, for example. The blood level notification device 1 is connected to, for example, an imaging device 2 and a blood reservoir tank 3, as well as a soft reservoir 3a, a centrifugal pump 3b, and an oxygen supply device 3c. The blood level notification device 1 may also be connected to other terminals 5 or servers 6 via a communication network 4, for example. If there are multiple blood reservoir tanks 3 to be notified, the blood level notification device 1 may be configured to connect to multiple imaging devices 2.

[0033] The blood storage level notification system 100 acquires image data by photographing the blood storage reservoir tank 3 during surgery via the imaging device 2. The blood storage level notification system 100 acquires, for example, the brightness values ​​of each element R (red), G (green), and B (blue) in the RGB color space included in the image data. The blood storage level notification system 100 stores the acquired image data in a storage device such as the hard disk of the imaging device 2, the blood storage level notification device 1, or another terminal 5 or server 6. When performing calculation processing such as image data correction, the blood storage level notification system 100 may load the image data into the working memory of each device performing the operation and execute the processing.

[0034] The blood storage level notification system 100 sequentially scans multiple block pixels included in the image data of the blood storage reservoir tank 3 that constitute the acquired image data. Based on the scanning results, the blood storage level notification system 100 may appropriately correct the image data by converting its brightness, density, luminance, and color space based on the RGB values ​​included in each block pixel.

[0035] The blood storage level notification system 100 may use the blood storage level notification device 1 as a processing device for correcting image data, for example, or as an evaluation device for evaluating image data of the blood storage reservoir tank 3 acquired by the imaging device 2.

[0036] The blood storage level notification system 100 images the blood storage reservoir tank 3 via the imaging device 2, performs color tone change processing in the HSV color space of the blood (described later), and detects the blood storage level based on the set ROI, and after these processes, detects image data indicating the corrected and transformed blood storage level.

[0037] The blood storage level notification system 100 refers to a reference database, described later, and generates evaluation results for the image data. The blood storage level notification system 100 outputs, for example, image information corresponding to the evaluation results. The image data may include multiple pieces of data, such as numerical data indicating the RGB values ​​contained in block pixels, brightness data indicating the brightness of the image data, location data regarding the place and environment where the image was taken, shooting data such as shooting conditions and settings, and various other data associated with the captured image data.

[0038] Image data consists of multiple block pixels that make up an image, and formats such as bitmap or JPEG are used. Image data is represented as a series or collection of dots called colored dots, which are divided into a grid of fine dots, and each of these dots is assigned information such as color and brightness, thus representing the entire image.

[0039] The image data may include, for example, identification data to identify the blood reservoir tank 3 to be notified when the blood reservoir tank 3 is imaged, imaging equipment data, location data, time data, operator data, etc. Various types of data may be captured simultaneously, or, for example, individually captured image data may be linked and recorded accordingly.

[0040] The image data may include, for example, positional data and color data captured from multiple different directions based on information indicating the shape of the blood reservoir tank 3, for the image of the blood reservoir tank 3 to be notified. In this case, the accuracy of the relative positional relationship for each positional data can be improved.

[0041] The blood storage level notification system 100 sequentially scans the uncorrected RGB values ​​contained in multiple block pixels that make up the image data, based on the acquired image data. The blood storage level notification system 100 may also binarize the acquired image data and set the area ratio of light and dark in the image data. The blood storage level notification system 100 refers to the acquired image data of the blood storage reservoir tank 3 and converts the change in the number of pixels in the vertical direction (up and down) of the image information, mainly red (R), which is the component color of blood, into numerical data.

[0042] The blood storage level notification system 100 detects the blood storage level based, for example, on the color change of the blood in the HSV color space of the image data and a set ROI (region of interest), and generates and outputs a numerical data graph showing the hemodynamic fluctuations over time.

[0043] The blood storage level notification system 100 may also, for example, analyze image data (described later) using an image recognition AI to detect the blood storage level in the blood storage reservoir tank 3. The image recognition AI may be stored in, for example, the blood storage level notification device 1, the imaging device 2, the terminal 5, or the server 6, and may perform image analysis processing on each device in conjunction with the activation of the imaging process of the blood storage reservoir tank 3.

[0044] The blood storage level notification system 100 detects the blood storage level using image recognition AI, then compares the temporal changes in the blood storage level with past blood storage level data using a machine learning model described later, and predicts the future blood storage level. Furthermore, based on the prediction by the machine learning model, for example, the blood storage level notification system 100 sets a timing for when the blood storage reservoir tank 3 reaches a predetermined threshold, and provides notification corresponding to the prediction a predetermined time before the set timing.

[0045] (Blood storage level notification device 1) Next, an example of the blood storage level notification device 1 in this embodiment will be described with reference to Figure 2. Figure 2(a) is a schematic diagram showing an example of the configuration of the blood storage level notification device 1 in this embodiment, and Figure 2(b) is a schematic diagram showing an example of the function of the blood storage level notification device 1 in this embodiment.

[0046] As the blood storage level notification device 1, an electronic device such as a personal computer (PC) may be used, or an electronic device such as a smartphone, tablet terminal, wearable terminal, IoT (Internet of Things) device, or a single-board computer such as Raspberry Pi® may be used. The blood storage level notification device 1 comprises a housing 10, a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, a storage unit 104, and I / F 105~107, as shown in Figure 2(a). Each component 101~107 is connected by an internal bus 110.

[0047] The CPU 101 controls the entire blood level notification device 1. The ROM 102 stores the operating code for the CPU 101. The RAM 103 is a work area used when the CPU 101 is operating. The storage unit 104 stores various information such as image data and reference databases. As the storage unit 104, a data storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive) may be used. The blood level notification device 1 may also have a GPU (Graphics Processing Unit), which is not shown in the figure. Having a GPU enables faster computation processing than usual.

[0048] I / F105 is an interface for sending and receiving various types of information with the imaging device 2, and may also be an interface for sending and receiving various types of information with other terminals 5 or servers 6 via a communication network 4 such as the Internet.

[0049] I / F106 is an interface for sending and receiving information with the input unit 108. For example, a keyboard is used as the input unit 108, and the administrator of the blood level notification device 1 inputs various information or control commands for the blood level notification device 1 via the input unit 108. I / F107 is an interface for sending and receiving various information with the output unit 109. The output unit 109 outputs various information stored in the storage unit 104, or the processing status of the blood level notification device 1. A display is used as the output unit 109, and may be a touch panel, for example.

[0050] <Image Recognition AI> Image recognition AI utilizes technologies such as the well-known Convolutional Neural Network (CNN). For example, the image recognition AI is used to detect the blood level in a blood reservoir tank 3. It analyzes image data acquired by the imaging device 2 and evaluates the blood level in the reservoir tank 3 in real time. The image recognition AI identifies the liquid level by performing HSV (Hue-Saturation-Brightness) conversion and analyzing changes in blood color. It also sets a Region of Interest (ROI) to focus analysis on the blood portion of the image data, suppressing the detection of differences and changes.

[0051] Image recognition AI can be combined with, for example, known machine learning and deep learning techniques to compare with past image data and predict changes in blood volume with high accuracy. Furthermore, it may be combined with, for example, optical sensors to minimize errors.

[0052] <Machine learning model (reference database)> Machine learning models are stored in the storage unit 104 as various learning models, for example, as a reference database. The learning models have a correlation between previously acquired historical image data and reference information associated with that historical image data. The reference database may store, for example, historical image data and reference information. The learning models are constructed using machine learning with multiple sets of training data, for example, historical image data and reference information as a set of training data. As a learning method, deep learning such as a convolutional neural network may be used.

[0053] In this case, for example, correlation indicates the degree of connection between many-to-many pieces of information. Correlation is updated as needed during the machine learning process. Therefore, the learning model represents a classifier with optimized correlation (e.g., a function) based on, for example, past image data and reference information. Thus, evaluation results for image data are generated using a learning model constructed based on all the results of evaluating the imaging state of the image data to be reported in the past. This makes it possible to generate optimal evaluation results even when the image data to be reported has a complex imaging state.

[0054] Furthermore, the system can quantitatively generate optimal evaluation results even when the image data is identical or similar to past image data, or even when it is dissimilar. By improving the generalization ability during machine learning, the blood storage level in the blood storage reservoir tank 3 can be evaluated more accurately, and the accuracy of notifications according to hemodynamics can be improved.

[0055] Furthermore, the correlation may have multiple correlation degrees, for example, indicating the degree of connection between each past image data and each reference piece of information. The correlation degrees can be associated with weight variables, for example, when the learning model is constructed using a neural network.

[0056] Past image data contains the same type of information as the image data described above. Past image data may include, for example, multiple image data acquired when the image data to be reported was evaluated in the past.

[0057] The reference information is linked to past image data and indicates information about the imaging status of the image of the blood reservoir tank 3 that is subject to correction notification. The reference information may include, for example, binarization information indicating the type and state of the image to be notified, a default image indicating the optimal value of the light-to-dark area ratio, division information of the number of pixels of block pixels that make up the image to be notified, numerical images of the divided block pixels, calculated images, or default images indicating the optimal value of the numerical image after processing.

[0058] Furthermore, the reference information may include, for example, images for each wavelength range showing the degree of brightness of the block pixels constituting the image to be reported, a default image showing the optimal value for each wavelength range, and color correction of the color type and state of the HSV value converted based on the RGB values ​​of the block pixels constituting the image to be reported.

[0059] The reference information may also include, for example, a default image and range showing the optimal ROI value, the type and state of edge intensity of the skeleton (edge ​​enhancement) representing the frequency distribution of RGB values ​​of block pixels constituting the image to be reported, and a default image showing the optimal value of the state. The specific content and settings included in the reference information, as well as the default images for various optimal values, can be added or set as needed.

[0060] Furthermore, correlation may indicate the degree of connection between past image data and reference information, as shown in Figure 4, for example. In this case, by using correlation, the degree of relationship between multiple data points contained in the reference information (in Figure 4, "Reference A" to "Reference C") can be linked and stored for each of the multiple data points contained in the past image data (in Figure 4, "Image Data A" to "Image Data C"). Therefore, for example, through correlation, it is possible to link multiple data points contained in the reference information to a single data point contained in the past image data, thereby enabling the generation of multifaceted evaluation results.

[0061] The correlation has multiple degrees of correlation, which link multiple data points included in past image data (e.g., pixel data, imaging time, HSV, ROI, etc.) with multiple data points included in reference information (e.g., blood storage level, etc.). The degree of correlation may be expressed in three or more stages, such as a percentage, a 10-point scale, or a 5-point scale, or it may be expressed by line characteristics (e.g., line thickness, etc.).

[0062] Correlation indicates, for example, that "Data A" included in past image data has a correlation degree of AA (75%) with "Reference A" included in reference information, and a correlation degree of AB (50%) with "Reference B" included in reference information. Also, for example, "Data B" included in past image data has a correlation degree of BA (35%) with "Reference A" included in reference information. In other words, "correlation degree" indicates the degree of connection between each data; for example, a higher correlation degree indicates a stronger connection between each data. Note that when constructing a learning model using the machine learning method described above, the correlation degree may be set to have three or more levels of correlation.

[0063] Furthermore, past blood storage level data may be stored in a reference database, for example, as shown in Figure 5, if it has correlations between multiple data points included in past blood storage level data (e.g., blood storage time, blood storage level / blood volume ratio, etc.) and multiple data points included in reference information linked to previously set timings (e.g., timing, blood storage time / predicted time, blood storage increase / decrease ratio, etc.). In this case, the degree of correlation is calculated based on the relationship between the combination of past blood storage level data and past timing (hemodynamics) and the reference information. In addition to the above, past blood storage level data may also be stored in the reference database in combination with past prediction results and hemodynamics.

[0064] For example, the combination of "Blood Storage Level Data A" included in past blood storage level data and "Reference AA" included in past timing A shows a correlation of AA (80%) with "Reference AA" and a correlation of AB (65%) with "Reference BB". In this case, past blood storage level data and past timing (hemodynamic) data can be stored independently. Therefore, it is possible to improve accuracy and expand the range of options when generating evaluation results.

[0065] Figure 2(b) is a schematic diagram showing an example of the functions of the blood storage level notification device 1. The blood storage level notification device 1 comprises an image acquisition unit 11, a blood storage level detection unit 12, a prediction unit 13, a timing setting unit 14, and a notification unit 15, and may also have, for example, an update unit 16. Note that each function shown in Figure 2(b) is realized by the CPU 101 executing a program stored in the storage unit 104, etc., using the RAM 103 as a work area, and may be controlled, for example, by artificial intelligence.

[0066] <<Image acquisition unit 11>> The image acquisition unit 11 acquires image data by, for example, photographing the blood reservoir tank 3 during surgery. The image acquisition unit 11 acquires the brightness values ​​of each element R (red), G (green), and B (blue) in the RGB color space included in the image data and stores them in a storage device such as the hard disk of the imaging device 2, the blood reservoir level notification device 1, or another terminal 5 or server 6.

[0067] The image acquisition unit 11 may, for example, load the image data into the working memory of each device performing the operation and execute the processing when performing calculations such as image data correction. The image acquisition unit 11 sequentially acquires multiple block pixels that constitute the image data. In addition to acquiring image data from the imaging device 2, the image acquisition unit 11 also acquires image data from, for example, a built-in imaging unit (not shown). The frequency and period at which the image acquisition unit 11 acquires image data are arbitrary.

[0068] The image acquisition unit 11 may be, for example, a digital camera, a digital video camera, or a known USB camera. The image acquisition unit 11 may be fixedly installed in a position where it can image the blood reservoir tank 3, or multiple units may be installed. The image acquisition unit 11 may transmit the captured image data to an external terminal such as another terminal 5 via, for example, the communication network 4 and the I / F 105.

[0069] <<Blood storage level detection unit 12>> The blood storage level detection unit 12 refers to the image data of the blood storage reservoir tank 3 acquired by the image acquisition unit 11 and converts the vertical (up and down) pixel count changes of the image information, mainly red (R), which is the component color of blood, into numerical data. The blood storage level detection unit 12 detects the blood storage level based on, for example, the color tone changes of the blood in the HSV color space of the image data and the set ROI (region of interest), and represents the detected blood storage level as numerical data showing its fluctuation over time.

[0070] The ROI is defined by the administrator, person in charge, operator, etc., as the range of the target to be detected. The defined ROI may be, for example, a specific numerical area on the front of the blood reservoir tank 3. The blood level detection unit 12 detects the blood level within the defined ROI and converts it into numerical data.

[0071] Furthermore, the blood storage level detection unit 12 generates a linear curve representing hemodynamics using the converted numerical data. The linear curve is a graph that represents the real-time temporal change of the blood storage level, for example, as shown in Figure 3, with the X-axis representing the time (seconds) of the change in the blood storage level in the blood storage reservoir tank 3 and the Y-axis representing the height (mm) of the blood storage level in the blood storage reservoir tank. The blood storage level detection unit 12, for example, uses the rate of change in the blood storage level in the blood storage reservoir tank 3 at an arbitrary point in time as the "slope: a" and the blood storage level as the "intercept: b", and the change in the numerical value of the pixel data as the level change, converts it into an approximate straight line of the form "f(x)=ax+b", and generates a graph that represents the real-time temporal change of the blood storage level.

[0072] The blood storage level detection unit 12 detects the blood storage level in intervals a to f based on the generated linear curve. The detected blood storage level may be displayed on the monitor of the blood storage level notification device 1, for example, by combining the generated linear curve with image data showing the filled portion of the blood storage reservoir tank 3. The blood storage level detection unit 12 may, for example, display the image data showing the filled portion in red on the monitor, and may also display the upper end of the blood storage level in the filled portion as a red linear image data.

[0073] Furthermore, the blood storage level detection unit 12 may use, for example, a generated linear curve or image data acquired by the image acquisition unit 11 as input data, analyze it with an image recognition AI, and detect the blood storage level in the blood storage reservoir tank 3. The image recognition AI may be, for example, a known image recognition AI, and may be used to detect the blood storage level using, for example, an image recognition AI (not shown) with multiple types of features. Multiple image recognition AIs may be appropriately selected in advance according to conditions such as the surgical procedure or the patient's condition.

[0074] The blood storage level detection unit 12 converts changes in the blood storage level into numerical data, for example, changes in the vertical (up and down) number of pixels in image information, mainly red, which is the color of blood components. If the ROI (Region of Interest) has a predetermined range, the blood storage level detection unit 12 also takes in information about the set range and detects the blood storage level targeting the range of interest set by the ROI (for example, the part of the front of the blood storage reservoir tank 3 that shows a specific numerical value).

[0075] The blood storage level detection unit 12 detects the blood storage level based on the color change in the HSV color space of the blood in the image data and the set ROI. It may also perform detection by changing the processing order, set the level in real time, change the setting conditions as appropriate, or, for example, set the process to be executed multiple times.

[0076] <<Prediction Unit 13>> The prediction unit 13 predicts future blood storage levels based, for example, on the real-time temporal changes in the blood storage level in the blood storage reservoir tank 3 detected by the blood storage level detection unit 12 (a graph showing the real-time temporal changes in the blood storage level).

[0077] The prediction unit 13 predicts the blood storage level based on a linear curve generated by the blood storage level detection unit 12. It uses an "intercept" indicating the blood storage level in the blood storage reservoir tank 3 at any given time (e.g., each interval a to f) and a "slope" indicating the rate of change in the blood storage level to predict whether the hemodynamics of the blood storage level will remain horizontal, decrease, or increase.

[0078] The prediction unit 13, for example, uses the numerical changes of block pixels in image data acquired in real time by the image acquisition unit 11 to predict future blood storage levels by going back several seconds to tens of seconds and using the pixel data acquired up to that point.

[0079] The prediction unit 13 predicts the value "f(x+n)" from the present to a set value (empty, full) using, for example, the aforementioned approximation line. This makes it possible to predict the hemodynamics of the blood storage level in the blood storage reservoir tank 3. The number of continuous values ​​to sample and the prediction range of the blood storage level (for example, how many seconds into the future) set by the prediction unit 13 when making predictions may be set arbitrarily.

[0080] The prediction unit 13 predicts, for example, the change in blood storage level in interval b (approximate line: y=-0.64x+200) from interval a (approximate line: y=202) to the inflection point interval b (approximate line: y=-0.64x+200), for example The timing of notification is predicted for a decrease of 4 pixels / second. In this case, the prediction unit 13 predicts that the timing of notification will be 22.6 seconds when the blood level in the blood reservoir tank 3 decreases from 1500 ml to 66.4 ml / second.

[0081] Furthermore, the prediction unit 13 inflects, for example, from interval b (approximation curve: y = -0.64x + 200) At the point interval c (approximation curve: y = -1.89x + 135), blood storage in interval c The timing of notification is predicted in response to a change in level (e.g., a decrease of 11.4 pixels / second). Based on this, the timing of notification is predicted. In this case, for example, if the blood level in the blood reservoir tank 3 is 500 ml, and the hemodynamic decrease is 66.4 ml / second, the prediction unit 13 predicts a notification timing of 7.6 seconds as the response grace period.

[0082] Furthermore, the prediction unit 13 may use a machine learning model to compare the temporal changes in the blood storage level in the blood storage reservoir tank in each interval a to f detected by the blood storage level detection unit 12 through image data analysis using image recognition AI with past blood storage level data (a graph showing the blood storage level as a real-time temporal change) and predict future blood storage levels.

[0083] Furthermore, if, for example, patient vital signs and surgical progress data are added, the prediction unit 13 uses the patient's vital signs, surgical progress data, and the skills of the responder as additional parameters to predict the corresponding future blood storage level.

[0084] If, for example, the patient's vital signs (good) and surgical progress data (e.g., blood volume: low, completion time: 1 hour) are input, the prediction unit 13 refers to a machine learning model and evaluates that the hemodynamics are trending towards stability (the temporal changes in the "intercept" of the blood level and the "slope" of the rate of change are small), and predicts the blood level as follows: "Reservoir level fluctuation: intervals a and f are normal (hemodynamics), level threshold: interval b is emptied late (hemodynamics are slow), interval c is emptied early (hemodynamics are fast), interval d is filled early (hemodynamics are fast), interval e is filled late (hemodynamics are slow)."

[0085] Furthermore, if additional parameters such as the patient's vital signs (e.g., require observation) and surgical progress data (e.g., blood volume: high, completion time: 3 hours) are input to the prediction unit 13, it may refer to a machine learning model and, if it evaluates that the hemodynamics are fluctuating (large temporal changes in the "intercept" of the blood storage level and the "slope" of the fluctuation rate) (intervals c and d), predict the corresponding blood storage level as "reservoir level fluctuation: sudden change (hemodynamic), level threshold: (time when it becomes empty: early, time when it becomes full: early)".

[0086] Furthermore, if additional information such as the responder's skill information (e.g., work history, job rank / role, department, qualifications held, etc.) is entered, the prediction unit 13 will predict the future blood storage level appropriate to the responder based on the skill information. The prediction unit 13 may also evaluate the responder's experience level (experience, judgment, etc.), feasibility of handling the case, etc., based on the skill information entered.

[0087] The prediction unit 13, for example, if the responder's skill information is "veteran (e.g., work experience: A (●● years or more), job rank / role: A (●● or more), department (●● department), qualifications held: A (●● items), etc.)", sets the corresponding blood reservoir level as "veteran / expert" and predicts "reservoir level fluctuation: able to respond to trouble (hemodynamic), level threshold: (time when it becomes empty: normal, time when it becomes full: normal)".

[0088] Furthermore, the prediction unit 13 may, for example, if the responder has little experience and their skill information is "beginner (e.g., work experience: C (less than ● years), job rank / role: C (less than ●●), department (●● department), qualifications held: C (● items), etc.)", make predictions for "younger / beginner" types by distinguishing them as "reservoir level fluctuation: can only respond when stable (hemodynamic), level threshold: (time when it becomes empty: early, time when it becomes full: early)".

[0089] The prediction unit 13 can appropriately make predictions according to the situation by inputting and setting additional parameters such as the patient's biological information or the responder's skill information based on the interval a to f of the linear curve generated by the blood storage level detection unit 12. This makes it possible to more accurately evaluate the blood storage level in the blood storage reservoir tank and improve the accuracy of notification according to hemodynamics.

[0090] <<Timing setting section 14>> The timing setting unit 14 sets the timing at which the blood reservoir tank 3 reaches a predetermined threshold, based, for example, on the temporal change in the future blood storage level predicted by the prediction unit 13. The timing setting unit 14 may, for example, set the numerical value of the timing (time) at which the blood reservoir tank 3 becomes empty as the first threshold when the tank tends to become empty (e.g., in intervals b and c), or it may set the numerical value of the timing (time) at which the tank becomes full as the second threshold when the tank tends to become full (e.g., in intervals d and e).

[0091] The timing set by the timing setting unit 14 may be configured to set multiple timings based on the prediction results predicted by the prediction unit 13. The timing setting unit 14 sets multiple notification timings based on the prediction results, for example, the future blood storage level predicted by the prediction unit 13, and further evaluates various additional parameters that have been input (patient's vital signs and surgical progress data, responder's skill information, etc.).

[0092] The timing setting unit 14 may, for example, assume that the parameters additionally input to the prediction unit 13 are "patient's biological information and surgical progress data," and that the predicted blood reservoir level resulting from these parameters is, for example, "reservoir level fluctuation: normal (hemodynamic), level threshold: interval b - time of emptying is late (hemodynamics are slow), interval c - time of emptying is early (hemodynamics are fast), interval d - time of fullness is early (hemodynamics are fast)." In this case, the timing setting unit 14 may evaluate that the hemodynamics are in a stable trend (interval a) and that the time of emptying and fullness are also in a late trend, and set the timing to "a little before the time of emptying / fullness (n seconds) (for example, 5 seconds before)."

[0093] Furthermore, the timing setting unit 14 assumes, for example, that the parameters additionally input to the prediction unit 13 are "skill information of the responder, etc.," and as a result, the prediction result of the blood reservoir level by the prediction unit 13 is, for example, that the responder is a "veteran," the blood reservoir level setting is "veteran expert," and "reservoir level fluctuation: can handle trouble (hemodynamic), level threshold: (time when it becomes empty: normal, time when it becomes full: normal)."

[0094] In this case, the timing setting unit 14 evaluates the system by setting "Veteran Expert" based on, for example, the skill information of the responder, specifically "Veteran". The timing setting unit 14 may evaluate, for example, from the generated linear curve that the hemodynamics are trending towards stability (interval a), and that the time when the system becomes empty (intervals b, c) and when it becomes full (d, e) is also trending towards being late, and set the timing for setting the notification as "Veteran" to "a little before the time when the system becomes empty / full (n seconds) (for example, 3 seconds before)".

[0095] Similarly, the timing setting unit 14, for example, if the response is for the aforementioned "junior / beginner" type, the prediction result is, for example, "Reservoir level fluctuation: Only possible when stable (hemodynamic), Level threshold: (Time when it becomes empty: early, Time when it becomes full: early)." In this case, the timing setting unit 14 sets and evaluates the response as a "junior / beginner" based on the response's skill information.

[0096] The timing setting unit 14 evaluates, for example, from the generated linear curve, that the hemodynamics are in a stable trend (interval a), and that the time when the tank becomes empty (intervals b, c) and when it becomes full (intervals d, e) is also in a late trend. However, the timing for setting the notification may be set to "a time significantly earlier than the time when the tank becomes empty or full (n seconds) (for example, 15 seconds earlier)" for "younger users / beginners."

[0097] The timing setting unit 14 may, for example, set the timing at which the blood reservoir tank 3 reaches a predetermined threshold on the graph showing the temporal changes in hemodynamics shown in Figure 3, and display the set timing result. The timing setting unit 14 may set a predicted timing based on, for example, the patient's vital signs and surgical progress data input as additional parameters by the prediction unit 13, or the skills information of the person in charge. Alternatively, it may monitor fluctuations in the blood storage level of the blood reservoir tank 3 in real time and set a timing for notification according to the monitoring results. The time, range, and location of the setting are arbitrary.

[0098] <<Hochi Department 15>> The notification unit 15 provides notification corresponding to the prediction a predetermined time before the timing. The notification unit 15 provides notification corresponding to the predicted hemodynamics a predetermined time before the timing set in the timing setting unit 14. For example, the notification unit 15 provides notification according to each set timing if the timing set in the timing setting unit 14 is normal (hemodynamics are stable: slope is flat) as shown in sections a and e, empty (hemodynamics are decreasing: slope is negative) as shown in sections b and c, full (hemodynamics are increasing: slope is positive) as shown in sections d and e, or if there are multiple other hemodynamic conditions.

[0099] The notification unit 15 may also provide notification information including the remaining time until the blood reservoir tank 3 is empty or full, and information regarding the current hemodynamic status. The notification unit 15 may also provide notification information including, for example, the remaining time until the blood reservoir tank 3 is empty or full (e.g., "10 seconds left"), and information regarding the current hemodynamic status (e.g., "OK", "Warning", "Danger").

[0100] Furthermore, the notification unit 15 may display, in addition to the timing set by the timing setting unit 14, image data of the blood reservoir tank 3 detected by, for example, the blood storage level detection unit 12. The form in which the notification unit 15 displays or notifies can be any form that corresponds to the blood storage level, and may include display color, display size, and even sound or alarm notification. The forms, combinations, and patterns of each are arbitrary.

[0101] The notification unit 15 outputs the blood storage level notification result. The notification unit 15 transmits the blood storage level notification result to the output unit 109 via the I / F 107, and also transmits the blood storage level notification result to other terminals 5, etc., via the I / F 105, for example. The notification unit 15 may also output data to display the blood storage level notification result to the output unit 109, etc.

[0102] <<Updated part 16>> The update unit 16 updates the reference database, for example. When the update unit 16 acquires a new relationship between past image data and reference information, it reflects the relationship in the correlation. For example, if an administrator determines the accuracy of the evaluation results based on the evaluation results generated by the blood storage level detection unit 12, etc., and the blood storage level notification device 1 acquires the determination result, the update unit 16 updates the correlation included in the reference database based on the determination result.

[0103] Furthermore, the blood storage level detection unit 12, prediction unit 13, timing setting unit 14, and notification unit 15 (hereinafter referred to as each processing unit 12 to 15) may, for example, refer to a reference database and generate evaluation results for each processing result. Each processing unit 12 to 15 may, for example, take processing data as input data, select the optimal reference information associated with the solution calculated based on the learning model, and generate evaluation results based on the optimal reference information.

[0104] Of the processing units 12 to 15, for example, the blood storage level detection unit 12, when referring to the reference database shown in Figures 4 and 5, selects data that is identical or similar to the data contained in the image data (for example, "Data A": let's call it the first data). As the first data, data that partially or completely matches the image data may be selected, or similar data may be selected. If the image data is represented by numerical values ​​such as a matrix, the range of numerical values ​​included in the selected first data may be set in advance.

[0105] The blood storage level detection unit 12 selects reference information associated with the selected first data, and the degree of correlation (first correlation) between the selected first data and the reference information, and generates an evaluation result based on the selected reference information and the first correlation. The first correlation may be selected from pre-established correlations, or the output for each process may be calculated by the prediction unit 13, the timing setting unit 14, etc.

[0106] For example, the blood storage level detection unit 12 selects data "Reference A" included in the reference information linked to the first data "Data A," and the first correlation degree (correlation degree AA) "75%" between "Data A" and "Reference A." Note that the reference information and the first correlation degree may include multiple data. In this case, in addition to "Reference A" and "75%" as described above, the unit may also select reference information "Reference B" linked to the first data "Data A," and the first correlation degree (correlation degree AB) "12%" between "Data A" and "Reference B," and generate an evaluation result based on "Reference A" and "75%," as well as "Reference B" and "12%."

[0107] The evaluation results may include image data. Furthermore, the first correlation index may be expressed in three or more stages, such as a percentage.

[0108] The blood storage level detection unit 12 uses format data, such as an output format, pre-stored in the storage unit 104, to generate an evaluation result that shows the selected reference information and the first correlation degree, etc., in a format (e.g., a string) that the user can understand. Note that the format settings for generating the evaluation result may be based on known technologies, for example.

[0109] The blood storage level detection unit 12 determines the content of the evaluation result, for example, based on the selected first correlation degree. For example, the blood storage level detection unit 12 may be set to generate the evaluation result based on reference information associated with a first correlation degree of "50%" or higher, and not reflect reference information associated with a first correlation degree of less than "50%" in the evaluation result. The judgment criteria based on the first correlation degree may be set in advance by, for example, an administrator, and the range of the threshold can be set arbitrarily. In addition, the blood storage level detection unit 12 may determine the content of the evaluation result based on, for example, the result of calculating a first correlation degree of 2 or more, or a comparison of a first correlation degree of 2 or more.

[0110] <<Storage section>> The memory unit retrieves various information, such as reference databases, stored in the storage unit 104 as needed. The memory unit stores various information acquired or generated by each of the components 11, 13-15 in the storage unit 104.

[0111] <Communication Network 4> The communication network 4 is, for example, the Internet network to which the blood donation level notification device 1, etc., is connected via a communication circuit. The communication network 4 may consist of a so-called optical fiber communication network. Alternatively, the communication network 4 may be a limited intranet wired communication network used within the organization of a company, facility, or hospital, or it may be implemented using a known communication network such as a wireless communication network.

[0112] <Other devices 5> Other terminals 5 include, for example, electronic devices similar to the blood level notification device 1. Other terminals 5 include, for example, a central control unit that can communicate with multiple blood level notification devices 1.

[0113] <Server 6> Server 6 stores, for example, the various types of information described above. Server 6 also stores, for example, various types of information sent via the communication network 4. Server 6 may store, for example, the same information as the storage unit 104, and may send and receive various types of information with the blood level notification device 1 via the communication network 4. In other words, the blood level notification device 1 may use server 6 instead of storage unit 104.

[0114] (Embodiment: Method for notifying blood storage level) Next, an example of a blood storage level notification method in this embodiment will be described. Figure 6 is a flowchart of an example of a blood storage level notification method. The blood storage level notification method can be implemented using the blood storage level notification system 100 or the blood storage level notification device 1 described above. The blood storage level notification method comprises an image acquisition step S110, a blood storage level detection step S120, a prediction step S130, a timing setting step S140, and a notification step S150.

[0115] <Image acquisition step S110> As shown in Figure 1, the image acquisition unit 11 photographs the blood reservoir tank 3 and acquires image data (image acquisition step S110). The image acquisition unit 11 may acquire image data of the front or multiple sides of the blood reservoir tank 3, or it may acquire image data of the blood level.

[0116] The image acquisition unit 11 may, for example, sequentially scan multiple block pixels constituting the image data in a pinpoint or wide-area manner according to various information regarding the shape and condition of the blood reservoir tank 3, the room in which the blood reservoir tank 3 is installed, and other environmental information, and acquire the target image data.

[0117] For example, the image acquisition unit 11 may acquire various types of image data from the imaging device 2, etc., or it may acquire image data in real time from a built-in imaging unit (not shown). The image acquisition unit 11 may acquire image data as still images or as moving images, in which case it will acquire image data for each image frame. The frequency and period of image data acquisition are arbitrary. The format and type of images and moving images are also arbitrary.

[0118] <Blood storage level detection step S120> The blood storage level detection unit 12 detects the blood storage level in the blood storage reservoir tank 3 (blood storage level detection step S120). The blood storage level detection unit 12 analyzes the image data acquired by the image acquisition unit 11, for example, and detects the blood storage level in the blood storage reservoir tank 3. The blood storage level detection unit 12 performs brightness, density, luminance, and color space conversions of the image data based on the RGB values ​​contained in the block pixels of the image data, for example.

[0119] The blood storage level detection unit 12 detects the red color component of blood based on data indicating the state of color tone changes in the HSV (Hue-Saturation-Brightness) color space, for example, from the blood image data acquired by the image acquisition unit 11. The blood storage level detection unit 12 also detects the range of blood storage levels within a pre-set ROI range. Furthermore, the blood storage level detection unit 12 generates and displays a graph showing the temporal changes in hemodynamics, for example, as shown in Figure 3, based on the detected blood storage level.

[0120] <Prediction step S130> The prediction unit 13 predicts future blood storage levels (prediction step S130). The prediction unit 13 predicts future increases or decreases in blood storage levels based on, for example, the temporal changes in blood storage levels detected by the blood storage level detection unit 12. The prediction unit 13 also predicts future blood storage levels based on, for example, the blood storage level detection unit 12, which analyzes image data using image recognition AI and detects the blood storage level in the blood storage reservoir tank 3. The prediction unit 13 then compares the temporal changes in blood storage levels with past blood storage level data using a machine learning model to predict future blood storage levels.

[0121] The prediction unit 13 predicts the future blood storage level as additional parameters when additional information is entered, such as the patient's vital signs (requires observation), surgical progress data (e.g., blood storage volume: large, completion time: 3 hours), or the skills of the person in charge. By entering additional parameters, the blood storage level in the blood storage reservoir tank 3 can be evaluated more accurately, and the accuracy of notifications according to hemodynamics can be improved.

[0122] <Timing setting step S140> The timing setting unit 14 sets a timing that will be a predetermined threshold (timing setting step S140). The timing setting unit 14 sets the timing at which the blood reservoir tank 3 will be below a predetermined threshold (time before emptying: numerical value, time prior) or above a predetermined threshold (time before full filling: numerical value, time prior) based on, for example, the temporal change in the future blood storage level predicted by the prediction unit 13.

[0123] Furthermore, the timing setting unit 14 may, for example, in the prediction unit 13, compare the temporal changes in the blood storage level with past blood storage level data using a machine learning model, and set the timing for when the blood storage reservoir tank 3 will fall below a predetermined threshold (time before emptying: numerical value) or exceed the threshold (time before fulling: numerical value).

[0124] The timing setting unit 14 may, for example, set the timing at which the blood reservoir tank 3 reaches a predetermined threshold on the graph showing the temporal changes in hemodynamics shown in Figure 3, and display the set timing result. The timing setting unit 14 may set the timing predicted based on, for example, the patient's vital signs and surgical progress data additionally input by the prediction unit 13, or the skills information of the person in charge. Alternatively, it may monitor the fluctuations in the blood storage level of the blood reservoir tank 3 in real time and set the timing according to the monitoring results. The time, range, and location of the setting are arbitrary.

[0125] <Hochi Step S150> The notification unit 15 provides notification corresponding to the prediction a predetermined time before the timing (notification step S150). The notification unit 15 provides notification corresponding to the predicted hemodynamics a predetermined time before the timing set in the timing setting unit 14. If there are multiple timings set in the timing setting unit 14, such as normal (hemodynamics are stable: slope is flat), when it becomes empty (hemodynamics are decreasing: slope is negative), or when it becomes full (hemodynamics are increasing: slope is positive), the notification unit 15 provides notification according to each timing.

[0126] The notification unit 15 may provide notification information including the remaining time until the blood reservoir tank 3 is empty or full, and information regarding the current hemodynamic status. The notification unit 15 may provide notification information including, for example, the remaining time until the blood reservoir tank 3 is empty or full (e.g., "10 seconds left"), and information regarding the current hemodynamic status (e.g., "OK", "Warning", "Danger", etc.).

[0127] Furthermore, the notification unit 15 may display, in addition to the timing set by the timing setting unit 14, image data of the blood reservoir tank 3 detected by, for example, the blood storage level detection unit 12. The form in which the notification unit 15 displays or notifies may be appropriate to the blood storage level, and may include display color, display size, and even sound or alarm notification. The form, combination, and pattern of each are arbitrary.

[0128] This may terminate the operation of the blood storage level notification method in this embodiment.

[0129] According to this embodiment, the prediction unit 13 compares the temporal change in the blood storage level with past blood storage level data using a machine learning model to predict the future blood storage level. Therefore, the timing setting unit 14 can set the timing at which the blood storage reservoir tank 3 reaches a predetermined threshold based on the prediction by the machine learning model. This makes it possible to evaluate the blood storage level in the blood storage reservoir tank 3 more accurately and improve the accuracy of notifications according to hemodynamics.

[0130] According to this embodiment, the timing setting unit 14 sets the timing at which the blood reservoir tank 3 reaches a predetermined threshold based on predictions made by a machine learning model. As a result, the notification unit 15 can provide notification corresponding to the prediction a predetermined time before the timing. This allows for a more accurate evaluation of the blood storage level in the blood reservoir tank 3 and improves the accuracy of notifications in accordance with hemodynamics.

[0131] Furthermore, according to this embodiment, the blood storage level detection unit 12 converts RGB values ​​into HSV values, which represent the color space of the color tone. It then detects the blood storage level based on the set ROI. Therefore, the blood storage level in the blood storage reservoir tank 3 can be evaluated more accurately. This improves the accuracy of notification according to hemodynamics.

[0132] Furthermore, according to this embodiment, the learning model is constructed by machine learning using past image data and reference information as training data. Therefore, even when setting unknown image data that is different from past image data, quantitative evaluation can be performed. This makes it possible to evaluate the blood storage level of the blood storage reservoir tank 3 more accurately and improve the accuracy of notification according to hemodynamics.

[0133] While embodiments of the present invention have been described, these embodiments are presented as examples only and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims of the invention and its equivalents. [Explanation of Symbols]

[0134] 1: Blood level notification device 2: Imaging device 3: Blood reservoir tank 4: Communication Network 5: Terminal 6: Server 10: Cabinet 11: Image acquisition unit 12: Blood storage level detection unit 13: Prediction Department 14: Timing setting section 15: Hochi Department 16: Update section 100: Blood Donation Level Notification System 101: CPU 102 :ROM 103: RAM 104: Preservation Department 105 :I / F 106: I / F 107: I / F 108: Input section 109: Output part 110: Internal bus S110: Image acquisition step S120: Blood storage level detection step S130: Prediction step S140: Timing setting step S150: Hochi Step

Claims

1. A blood storage level notification system that notifies the blood storage level of a blood storage reservoir tank, Image acquisition means for photographing the blood storage reservoir tank and acquiring image data, A blood storage level detection means analyzes the image data and detects the blood storage level in the blood storage reservoir tank, A prediction means for predicting future blood storage levels based on the temporal changes in the blood storage level, additionally inputted patient biometric information, and surgical progress data, A timing setting means that sets the timing at which the blood storage reservoir tank reaches a predetermined threshold based on the above prediction, A notification means that provides notification corresponding to the prediction at a predetermined time before the aforementioned timing. A blood storage level notification system characterized by comprising the following features.

2. The blood storage level detection means analyzes the image data using image recognition AI and detects the blood storage level in the blood storage reservoir tank. The prediction means compares the temporal change in the blood storage level with past blood storage level data using a machine learning model to predict future blood storage levels. The timing setting means sets the timing at which the blood storage reservoir tank reaches a predetermined threshold based on the prediction by the machine learning model. A blood storage level notification system according to claim 1, characterized by the above.

3. The blood storage level detection means detects the blood storage level as numerical data indicating the time-series fluctuation of the blood storage level in the blood storage reservoir tank, based on the color change of the blood in the HSV color space of the image data and the set ROI. A blood storage level notification system according to claim 1, characterized by the above.

4. The aforementioned prediction means uses the skills information of the responder as additional input. A blood storage level notification system according to claim 1, characterized by the above.

5. The notification means provides notification information including the remaining time until the blood reservoir tank is empty or full, and information regarding the current hemodynamic status. A blood storage level notification system according to claim 1, characterized by the above.

6. A blood reservoir level notification device that notifies the blood reservoir level in a reservoir tank, An image acquisition unit that photographs the aforementioned blood storage reservoir tank and acquires image data, A blood storage level detection unit analyzes the aforementioned image data and detects the blood storage level in the blood storage reservoir tank, A prediction unit predicts future blood storage levels based on the temporal changes in the blood storage level and additionally input patient biometric information and surgical progress data. A timing setting unit sets the timing at which the blood storage reservoir tank reaches a predetermined threshold based on the above prediction, A notification unit provides notification corresponding to the prediction at a predetermined time before the aforementioned timing. A blood storage level notification device characterized by comprising the following features.

7. The blood storage level detection unit analyzes the image data using image recognition AI and detects the blood storage level in the blood storage reservoir tank. The prediction unit compares the temporal changes in the blood storage level with past blood storage level data using a machine learning model to predict future blood storage levels. The timing setting unit sets the timing at which the blood storage reservoir tank reaches a predetermined threshold, based on the prediction by the machine learning model. A blood storage level notification device according to claim 6, characterized by the above.

8. A method for notifying the blood level of a blood storage reservoir tank, The image acquisition step involves taking a photograph of the blood storage reservoir tank and obtaining image data, A blood storage level detection step involves analyzing the image data and detecting the blood storage level in the blood storage reservoir tank, A prediction step that predicts future blood storage levels based on the temporal changes in the blood storage level and additionally entered patient biometric information and surgical progress data, A timing setting step is to set the timing at which the blood storage reservoir tank reaches a predetermined threshold based on the above prediction. A notification step in which notification corresponding to the prediction is given a predetermined time before the aforementioned timing. A method for notifying blood storage level, characterized by having the following features.

9. The blood storage level detection step involves analyzing the image data using image recognition AI to detect the blood storage level in the blood storage reservoir tank. The prediction step involves comparing the temporal changes in the blood storage level with past blood storage level data using a machine learning model to predict future blood storage levels. The timing setting step involves setting the timing at which the blood storage reservoir tank reaches a predetermined threshold, based on the prediction by the machine learning model. A method for notifying blood storage level according to claim 8, characterized by the above.

Images