JP7264468B2 - Diagnosis support system and program - Google Patents

Description

The present invention relates to a diagnostic support system and program.

"Hand tremor" is an important diagnostic item in diagnosing neurological diseases. However, since many tremors are caused by physiological reasons rather than illness, people often wonder whether they should see a doctor even if they have tremors in their hands. In addition, in order to determine whether tremor is pathological or not, a certain amount of experience as a neurologist is required, and the reality is that only a very limited number of doctors can make highly accurate diagnoses. Met.

Under such circumstances, there is a demand for a diagnostic support system that can quantitatively and simply determine whether or not tremor requires diagnosis and treatment. For example, Non-Patent Document 1 describes a system that allows a subject to draw a spiral with a stylus pen on a touch panel of a tablet computer, and analyzes whether the subject is a healthy person based on the drawing result. is disclosed.

Further, in Patent Document 1, means for reading pen drawings as positional information is attached to a general-purpose computer system, and the motion trajectory of the subject's drawing work is taken into the computer along with time-lapse information, thereby diagnosing cerebral and nervous system diseases. An assistance system is disclosed for assisting in diagnosing medical conditions. This support system compares and displays a parameter related to the drawing work of the subject and a standard numerical value of the parameter related to a healthy person or a numerical value representing phenomena/symptoms peculiar to cerebral/nervous system diseases.

Jonathan A. Sisti, Brandon Christophe, Audrey Rakovich Seville, Andrew L.A. Garton, Vivek P. Gupta, Alexander J. Bandin, Qiping Yu, and Seth L. Pullman, "Computerized Spiral Analysis Using the iPad", HHS Public Access Author manuscript J Neurosci Methods. Author manuscript; Published online 2016 Nov 10. doi: 10.1016/j.jneumeth.2016.11.004

Japanese Patent No. 4524054

The system disclosed in Non-Patent Literature 1 requires complicated work before introduction, such as preparing a tablet computer and installing an analysis program on the tablet computer. In addition, the spirals drawn by this system vary in size and shape even in healthy subjects, which hinders quantitative diagnosis. Furthermore, in this system, since the diagnosis is performed in consideration of the writing pressure and the drawing speed at the time of drawing, complicated calculations are required, so it is not suitable for simple determination. Due to this background, the actual situation is that the introduction of this system to clinical sites has not progressed.

In addition, the system disclosed in Patent Document 1 only compares and displays the parameters related to the drawing work of the subject with those specific to healthy subjects or brain and nervous system diseases, and the diagnosis of the disease itself is substantially The reality is that it is up to the doctor.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a diagnostic support system and program capable of more accurately and simply determining the symptoms of a subject.

In order to achieve the above object, the diagnostic support system according to the first aspect of the present invention comprises:
an imaging unit for imaging a sheet on which a reference line pattern is printed and a handwriting pattern drawn by a subject by tracing the reference line pattern using a writing instrument;
a determination unit that determines symptoms related to the brain or nerves of the subject based on imaging data captured by the imaging unit;
with
The imaging unit is mounted on a mobile terminal,
The determination unit is implemented in a server computer that can communicate with the mobile terminal.

In this case, the determination unit determines the symptoms of the subject by performing deep learning using imaging data for which the determination result of the symptoms of the subject is known as teacher data.
You can do it.

Further, the determination unit
Determining the symptoms of the subject using a convolutional neural network trained using the teacher data, with the imaging data as input and the result of determining the symptoms of the subject as output.
You can do it.

In the judgment result,
that the subject is likely to be healthy, that the subject is likely to be suffering from essential tremor, and that the subject is likely to be suffering from cerebellar disorder high, including at least one of
You can do it.

The determination unit is
calculating at least one of a total amount of error between the reference line pattern and the handwriting pattern and a difference in total length between the reference line pattern and the handwriting pattern as an index value based on the imaging data;
Determining the symptom of the subject based on the calculated index value;
You can do it.

The determination unit is
If the index value is smaller than the first threshold, it is determined that the subject is likely to be healthy,
If the index value is greater than or equal to the first threshold and less than the second threshold, determining that the subject is likely to be suffering from essential tremor,
If the index value is equal to or greater than the second threshold, it is determined that the subject is likely to be suffering from cerebellar disorder.
You can do it.

Alignment patterns are printed on the sheet at vertices of a virtual rectangle containing the reference line pattern,
The imaging data includes an image of the alignment pattern together with the image of the handwriting pattern,
a correction unit that corrects the imaging data based on the relative positional relationship of the image of the alignment pattern in the imaging data;
The determination unit determines the symptoms of the subject based on the imaging data corrected by the correction unit.
You can do it.

A first of the alignment patterns is printed at a first density when expressed in gray scale, and a remaining second alignment pattern is printed at a density lower than the first density. It is printed with a density of 2,
The reference line pattern is printed at the second density when expressed in grayscale,
The handwriting pattern is drawn with a third density between the first density and the second density when expressed in grayscale,
The correcting unit extracts the image of the handwriting pattern based on a change in luminance in a range from the luminance of the image of the first alignment pattern to the luminance of the image of the second alignment pattern, binarize the imaging data,
The determination unit determines whether the subject is healthy based on the imaging data binarized by the correction unit.
You can do it.

the baseline pattern is a spiral;
You can do it.

A program according to a second aspect of the present invention comprises
A computer that operates as a mobile terminal,
Functioning as an imaging unit for imaging a sheet on which a reference line pattern is printed and a handwriting pattern drawn by a subject by tracing the reference line pattern with a writing instrument is formed,
a server computer capable of communicating with the mobile terminal,
It functions as a determination unit that determines symptoms related to the brain or nerves of the subject based on imaging data captured by the imaging unit.

According to the present invention, the subject's symptoms can be determined simply by having a sheet and an object capable of capturing an image of the sheet at hand of the subject. Moreover, according to the present invention, handwriting patterns drawn by tracing a reference line pattern are used, so that the symptoms of a subject can be determined under uniform conditions for healthy subjects. Furthermore, according to the present invention, only the imaging data of the handwriting pattern is used to determine the symptoms of the subject. Therefore, the symptoms of the subject can be determined more accurately and easily.

1 is a schematic diagram showing the configuration of a diagnostic support system according to Embodiment 1 of the present invention; FIG. FIG. 4 is a diagram showing an example of a pattern printed on a sheet; (A) is a diagram showing an example of a handwriting pattern drawn by an able-bodied person. (B) is a diagram showing an example of a handwriting pattern drawn by a subject suffering from essential tremor. (C) is a diagram showing an example of a handwriting pattern drawn by a subject suffering from cerebellar disorder. (A) is a graph showing an error area between a reference line pattern and a handwriting pattern. (B) is a graph showing the ratio of the total length of the handwriting pattern to the total length of the reference line pattern. (A) is a diagram showing an example of imaging data before correction. (B) is a diagram showing an example of imaging data after trapezoidal correction. (C) is a diagram showing an example of imaging data after binarization. It is a schematic diagram which shows the structure which performs deep learning. It is a figure which shows the hardware constitutions of a portable terminal. It is a figure which shows the hardware constitutions of a server computer. 2 is a flow chart showing the flow of processing of the diagnosis support system of FIG. 1; FIG. 9 is a schematic diagram showing a flow of calculating an error between a reference line pattern and a handwriting pattern in a diagnosis support system according to Embodiment 2 of the present invention;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In all figures, the same reference numerals are attached to the same or corresponding components.

Embodiment 1
First, Embodiment 1 of the present invention will be described. As shown in FIG. 1, a sheet 2 is used in the diagnosis support system 1 according to the present embodiment. As shown in FIG. 2, the sheet 2 is a sheet on which a reference line pattern BL is printed. The reference line pattern BL serves as a reference for the handwriting pattern WL drawn by the subject E using the writing instrument F. In this embodiment, the reference line pattern BL is a spiral (spiral line).

Four alignment patterns P1, P2, P3, and P4 are printed on the sheet 2 in addition to the reference line pattern BL. The alignment patterns P1 to P4 are arranged at the vertices of a virtual rectangle containing the reference line pattern BL. The positional relationship between the alignment patterns P1 to P4 is known, and the positional relationship between the alignment patterns P1 to P4 and the reference line pattern BL is also known.

The subject E draws a spiral by tracing the reference line pattern BL using a writing tool F such as a red magic pen. Thereby, a handwriting pattern WL is formed on the sheet 2 . As shown in FIG. 3A, when the subject E is healthy, the handwriting pattern WL substantially matches the reference line pattern BL.

However, as shown in FIG. 3(B), when the subject E is suffering from essential tremor, the handwriting pattern WL deviates little by little from the baseline pattern BL, and its total length (distance) is long. tend to become Further, as shown in FIG. 3(C), when the subject E suffers from cerebellar disorder, the handwriting pattern WL tends to deviate significantly from the reference line pattern BL and the overall length tends to be significantly longer. be.

In this manner, the handwriting pattern WL changes depending on whether the subject E is healthy. This change is represented by a deviation from the reference line pattern BL.

As an index value representing the deviation from the reference line pattern BL, for example, there is the area of the portion surrounded by the reference line pattern BL and the handwriting pattern WL. This is called the error area. The area of error was measured in healthy subjects (Normal; 22), essential tremor (ET) (13), and cerebellar disorder (CD) (16). 4(A). As shown in FIG. 4(A), when the subject E is healthy, the error area is 500 to 950 mm ² (the first, second, and third quartiles are 650, 700, and 750 mm ² ). In addition, when subject E suffers from essential tremor (ET), the error area is 500 to 1500 mm ² (800, 1050, 1200 mm ² for the first, second and third quartiles). ). Further, when subject E suffers from cerebellar disorder (CD), the error area is 900 to 3000 mm ² (first, second and third quartiles are 1000, 1300 and 1900 ^{mm 2} ). distributed. A significant difference was observed at the 5% level between the healthy subject and the essential tremor, and a significant difference was observed at the 1% level between the healthy subject and the cerebellar disorder and between the essential tremor and the cerebellar disorder.

As an index value representing the deviation from the reference line pattern BL, for example, there is a difference between the total length of the reference line pattern BL and the total length of the handwriting pattern WL. This difference can be expressed by the ratio of the total length of the reference line pattern BL and the total length of the handwriting pattern WL, that is, the total handwriting length [%]. When the total length of handwriting [%] was actually measured in healthy subjects (Normal; 22), essential tremor (ET) patients (13), and cerebellar disorder (CD) patients (16), the data was accumulated. The results are shown in FIG. 4(B). As shown in FIG. 4(B), when the subject E is a healthy subject, the handwriting total length is approximately 98 to 103% (the first, second, and third quartiles are about 101%). When subject E suffers from essential tremor (ET), the total length of the handwriting is 98 to 110% (first, second, third quartiles are 103, 104, 109%). In addition, when the subject E suffers from cerebellar disorder (CD), the handwriting total length is 100 to 145% (first, second, third quartile 105, 110, 120%) It's becoming A significant difference was observed at the 1% level between the healthy subject and the essential tremor, and a significant difference was observed at the 5% level between the essential tremor and the cerebellar disorder.

By analyzing the deviation of the handwriting pattern WL from the reference line pattern BL (by using the index value of the error area or the total length of handwriting [%]), the accumulation result of these data can be obtained by examining the subject E. This suggests that it is possible to determine whether or not a person is a patient, whether or not they are suffering from essential tremor, and whether or not they are suffering from cerebellar disorder.

As shown in FIG. 1, a diagnosis support system 1 is constructed by combining a portable terminal 10 and a server computer 20 to determine the symptoms of a subject E. As shown in FIG. The mobile terminal 10 is, for example, a smart phone, but is not limited to this, and may be a terminal (for example, a digital camera or a tablet computer) having an imaging function and a communication function. The mobile terminal 10 and the server computer 20 are connected via a network 15 such as the Internet, and are communicable with each other.

As shown in FIG. 1, the portable terminal 10 includes an imaging unit 10a, a display unit 10b, and a transmitting/receiving unit 10c as constituent elements of the diagnosis support system 1. FIG. The server computer 20 also includes a correction unit 20a, a determination unit 20b, and a transmission/reception unit 20c as constituent elements of the diagnosis support system 1. FIG.

The imaging unit 10a mounted on the mobile terminal 10 images the sheet 2 on which the handwriting pattern WL drawn by the subject E using the writing instrument F by tracing the reference line pattern BL is formed.

The display unit 10b mounted on the mobile terminal 10 displays the determination result 3. FIG.

The transmitting/receiving unit 10c mounted on the mobile terminal 10 transmits/receives data to/from the server computer 20. FIG. Specifically, the transmitting/receiving section 10c transmits imaged data 2a, which is the imaged result of the sheet 2, to the server computer 20 via the network 15. FIG. Furthermore, the transmitting/receiving unit 10 c receives the determination result 3 transmitted from the server computer 20 via the network 15 .

A correction unit 20 a mounted on the server computer 20 corrects the imaging data 2 a sent to the server computer 20 . This correction is performed in the following two stages.

(A) Tilt Correction Depending on the positional relationship between the sheet 2 and the portable terminal 10 at the time of photographing, the captured image data 2a of the sheet 2 has a trapezoidal shape or a parallelogram shape as shown in FIG. 5A. The imaging data 2a includes images P1a, P2a, P3a, and P4a of the alignment patterns P1, P2, P3, and P4 together with the image WLa of the handwriting pattern WL. Based on the relative positional relationship of the images P1a to P4a of the alignment patterns P1 to P4 in the imaged data 2a, the correction unit 20a converts the imaged data 2a shown in FIG. Correct (tilt correction) to data 2b. This correction is performed so that the positions of the images P1a, P2a, P3a, and P4a of the four alignment patterns P1 to P4 in the imaging data 2a are changed from the designed images P1b, P2b, P3b, and P4b of the alignment patterns P1 to P4. This is done by changing the position of each pixel of the imaging data 2a so as to match the position. As a result, the image BLa of the reference line pattern BL and the image WLa of the handwriting pattern WL of the imaging data 2a are converted into the image BLb of the reference line pattern BL and the image WLb of the handwriting pattern WL of the imaging data 2b.

(B) Binarization of Imaging Data Whether or not the subject E is healthy appears in the handwriting pattern WL. Therefore, the correction unit 20a erases the image BLb of the reference line pattern BL and extracts the image WLb of the handwriting pattern WL to obtain image data 2c in which the image WLc of the handwriting pattern WL is binarized.

Here, as shown in FIG. 2, among the four alignment patterns P1 to P4, the alignment patterns (first alignment patterns) P1 to P3 have a first density when expressed in grayscale. (e.g. black), and the remaining alignment pattern (second alignment pattern) P4 is printed at a second density lower than the first density (a color between white and black, e.g. light gray). ) is printed. Also, the reference line pattern BL is printed in the same second density (light gray) as the alignment pattern P4 when expressed in grayscale.

Furthermore, the handwriting pattern WL is drawn with a writing tool F (red magic pen) that has a third density between the first density and the second density when expressed in grayscale. Therefore, the reference line pattern BL and the handwriting pattern WL can be distinguished.

The correcting unit 20a extracts the image WLc of the handwriting pattern WL based on the change in luminance in the range from the luminance of the images of the alignment patterns P1 to P3 to the luminance of the image of the alignment pattern P4, while extracting the image WLc of the handwriting pattern WL. ) is binarized to obtain image data 2c shown in FIG. 5(C). In other words, the imaging data 2b is binarized image data in which the image WLc of the handwriting pattern WL is black and the other portions are white.

The determination unit 20b mounted on the server computer 20 has a high possibility that the subject E is a healthy person based on the imaging data 2c imaged by the imaging unit 10a, keystone-corrected by the correction unit 20a, and binarized. or likely to have essential tremor, or likely to have cerebellar disorder. In the present embodiment, the determination unit 20b performs deep learning using imaging data in which the determination result of the symptoms of the subject E is known (diagnosis results of the symptoms are confirmed) as teacher data. The symptoms of subject E are determined.

The determination unit 20b receives the imaging data 2c as an input and outputs a diagnosis result, and uses a convolutional neural network 21 learned using teacher data to determine whether or not the subject E needs to be diagnosed. judge. An example of this convolutional neural network 21 is shown in FIG.

As shown in FIG. 6, the convolutional neural network 21 is composed of three layers: an input layer, an intermediate layer, and an output layer. The input layer is provided by the number of pixels P(m, n) of the imaging data 2c, and the value of each pixel P(m, n), 0 or 1, is input. The number of pixels P(m, n) is M×N (64×64). In the output layer, three perceptrons are provided for determining that the possibility of a healthy subject is high, that the possibility of essential tremor is high, and that the possibility of cerebellar disorder is high. there is In the output layer, the output of any perceptron will be 1 and the rest will be 0.

When the value of each pixel P (m, n) of the imaging data is input to the input layer, the output from the input layer is transmitted to the intermediate layer, and the output of the intermediate layer is transmitted to the output layer. It is output whether there is a high possibility that the subject is a healthy subject, that the subject is likely to have essential tremor, or that the subject is likely to have a cerebellar disorder. That is, the determination result 3 indicates whether the subject E is likely to be healthy, whether the subject E is likely to be suffering from essential tremor, or whether the subject E is likely to have a cerebellar disorder. Either there is a high probability of being affected, or

The parameter adjustment unit 22 uses the teacher data 23 to make the convolutional neural network 21 learn. As the teacher data 23, for example, the imaging data 2c used to calculate the graphs of FIGS. The associated one is used. As a method of deep learning, for example, a function representing the error between the output of the output layer and the teacher data 23 is set as an error function, the error function is back-propagated toward the input layer, and the error function regarding the parameters of each layer Error backpropagation can be used to find the gradient of .

A transmitting/receiving unit 20c mounted on the server computer 20 transmits/receives data to/from the mobile terminal 10. FIG. Specifically, the transmission/reception unit 20c receives the imaging data 2a sent from the mobile terminal 10 via the network 15. FIG. Further, the transmitting/receiving unit 20c sends the determination result 3 of the determination unit 20b to the mobile terminal 10 via the network 15. FIG.

As shown in FIG. 7, the computer that operates as the mobile terminal 10 has a control unit 31, a main storage unit 32, an external storage unit 33, a camera 34, a touch panel 35, and a wireless communication interface 36 as hardware configurations. The main storage unit 32 , the external storage unit 33 , the camera 34 , the touch panel 35 and the wireless communication interface 36 are all connected to the control unit 31 via the internal bus 30 .

The control unit 31 is composed of a CPU (Central Processing Unit) and the like. This CPU executes the processing of the imaging unit 10a, the display unit 10b, and the transmission/reception unit 10c according to the program 39 stored in the external storage unit 33, so that the hardware and software work together as shown in FIG. The functions of the mobile terminal 10 are realized.

The main storage unit 32 is composed of a RAM (Random-Access Memory) and the like. A program 39 stored in the external storage unit 33 is loaded into the main storage unit 32 . In addition, the main storage unit 32 is used as a work area (temporary storage area for data) for the control unit 31 .

The external storage unit 33 is composed of a non-volatile memory such as a flash memory, hard disk, DVD-RAM (Digital Versatile Disc Random-Access Memory), DVD-RW (Digital Versatile Disc ReWritable). A program 39 to be executed by the control unit 31 is stored in advance in the external storage unit 33 . The external storage unit 33 also supplies the control unit 31 with data used in executing the program 39 according to instructions from the control unit 31 and stores the data supplied from the control unit 31 .

The camera 34 is built into the mobile terminal 10 . The camera 34 has an optical system and an imaging device. The imaging data 2a is obtained by receiving the light through the optical system with the imaging element. The imaging data 2 a is stored in the external storage unit 33 .

The touch panel 35 is a display screen on which operation input is possible. An operation input to the touch panel 35 is sent to the control unit 31, and the control unit 31 performs processing according to the operation input. According to this processing, the control unit 31 causes the display screen to display an image.

The wireless communication interface 36 is an interface circuit that wirelessly connects with a nearby WiFi router connected to the network 15, for example. Data communication with the server computer 20 via the network 15 is performed via this wireless communication interface 36 .

In the portable terminal 10, the control unit 31 executes the program 39 to realize the functions of the imaging unit 10a, the display unit 10b, and the transmission/reception unit 10c. Specifically, the functions of the imaging unit 10a are realized by the control unit 31, the main storage unit 32, the external storage unit 33, the camera 34, and the touch panel 35. The function of the display unit 10b is realized by the control unit 31, the main storage unit 32, the external storage unit 33, and the wireless communication interface 36, and the function of the transmission/reception unit 10c is realized.

As shown in FIG. 8, the computer operating as the server computer 20 has a control section 41, a main storage section 42, an external storage section 43, and a communication interface 44 as a hardware configuration. The main storage unit 42 , the external storage unit 43 and the communication interface 44 are all connected to the control unit 41 via the internal bus 40 .

The control unit 41 is composed of a CPU (Central Processing Unit) and the like. This CPU executes the processing of the correcting unit 20a, the determining unit 20b, and the transmitting/receiving unit 20c according to the program 49 stored in the external storage unit 43, whereby the hardware and software work together as shown in FIG. The functions of the server computer 20 are implemented.

The main storage unit 42 is composed of a RAM (Random-Access Memory) or the like. A program 49 stored in the external storage unit 43 is loaded into the main storage unit 42 . In addition, the main storage unit 42 is used as a work area (temporary storage area for data) for the control unit 41 .

The external storage unit 43 is composed of a non-volatile memory such as a flash memory, hard disk, DVD-RAM (Digital Versatile Disc Random-Access Memory), DVD-RW (Digital Versatile Disc ReWritable). A program 49 to be executed by the control unit 41 is stored in advance in the external storage unit 43 . In addition, the external storage unit 43 supplies data used in executing the program 49 to the control unit 41 according to instructions from the control unit 41 and stores the data supplied from the control unit 41 .

A communication interface 44 is a communication interface circuit for performing data communication via the network 15 .

In the server computer 20, the control section 41 executes the program 49 to realize the functions of the correction section 20a and the determination section 20b. Specifically, the functions of the correction unit 20a and the determination unit 20b are realized by the control unit 41, the main storage unit 42, and the external storage unit 43, and the functions of the control unit 41, the main storage unit 42, the external storage unit 43, and the communication interface 44 are realized. The function of the transmitting/receiving section 20c is realized.

Next, operation of the diagnostic support system 1 according to the embodiment of the present invention will be described.

As shown in FIG. 9, after the subject E draws the handwriting pattern WL on the sheet 2, the application of the diagnosis support system 1 is activated in the portable terminal 10, and the imaging unit 10a images the sheet 2 (step S11). ). Specifically, when the imaging button displayed on the touch panel 35 is pressed while the surface of the sheet 2 on which the handwriting pattern WL is drawn is within the imaging field, the control unit 31 drives the camera 34. Then, the sheet 2 is imaged. The control unit 31 stores the imaging data 2 a of the sheet 2 in the external storage unit 33 .

Subsequently, the transmission/reception unit 10c of the mobile terminal 10 transmits the imaging data 2a to the server computer 20 (step S12). Specifically, the control unit 31 connects to the network 15 via the wireless communication interface 36 . Subsequently, the control section 31 reads out the imaging data 2 a from the external storage section 33 and transmits it to the server computer 20 via the network 15 .

Subsequently, the transmitting/receiving unit 10c of the mobile terminal 10 waits for reception of the determination result 3 (step S13; No). Specifically, the control unit 31 waits for the determination result 3 to be sent from the server computer 20 via the wireless communication interface 36 .

On the other hand, in the server computer 20, the transmitting/receiving section 20c waits until the imaging data 2a is received from the mobile terminal 10, that is, waits for reception (step S21; No).

When the imaging data 2a is received from the mobile terminal 10 (step S21; Yes), the correction unit 20a of the server computer 20 corrects the imaging data 2a (step S22). Specifically, the control unit 41 reads the imaging data 2a received from the mobile terminal 10, converts the imaging data 2a shown in FIG. 5A into the imaging data 2b shown in FIG. 2b is converted into imaging data 2c shown in FIG. 5(C).

Further, the determination unit 20b of the server computer 20 performs determination processing (step S23). Specifically, as shown in FIG. 6, the control unit 41 inputs the imaging data 2c to the convolutional neural network 21 and obtains the output as the determination result. The determination result is stored in the external storage unit 43 .

Subsequently, the transmission/reception unit 20c of the server computer 20 transmits the determination result 3 to the mobile terminal 10 (step S24). Specifically, the control unit 41 connects to the network 15 via the communication interface 44 , reads the determination result 3 from the external storage unit 43 , and transmits it to the mobile terminal 10 .

When the transmitting/receiving unit 10c of the mobile terminal 10 receives the determination result 3 from the server computer 20 (step S13; Yes), the display unit 10b of the mobile terminal 10 displays the determination result (step S14). Specifically, the control unit 31 receives the determination result 3 via the wireless communication interface 36 and displays the received determination result 3 on the touch panel 35 . Looking at the determination result 3 displayed on the touch panel 35, there is a high possibility that the subject E is healthy, or that the subject E is likely to be suffering from essential tremor, or is suffering from cerebellar disorder. You can know what is likely.

Embodiment 2
Next, Embodiment 2 of the present invention will be described. The configuration of the diagnostic support system 1 according to the present embodiment is the same as the configuration of the diagnostic support system 1 according to the first embodiment shown in FIG. The diagnosis support system 1 according to the present embodiment differs from the determination section 20b of the first embodiment in the operation of the determination section 20b.

In the present embodiment, the determination unit 20b determines at least one of the total amount of error between the reference line pattern BL and the handwriting pattern WL and the difference in the total length between the reference line pattern BL and the handwriting pattern WL, based on the imaging data 2a. is calculated as an index value. Furthermore, the determination unit 20b determines the symptoms of the subject E based on these calculated index values.

First, a method of calculating the total amount of error between the reference line pattern BL and the handwriting pattern WL will be described. As shown in FIG. 10, points h _k (k=0 to N) are defined on the reference line pattern BL at minute intervals dL. Here, in the reference line pattern BL, let the origin O be the edge of the center of the spiral. When a straight line passing through the origin O and each point _hk is drawn, if the reference line pattern BL and the handwriting pattern WL are deviated, that point becomes a point _{Wk different from the point hk .} The determination unit 20b calculates the distance between the point _hk and the point _Wk as an error. The determining unit 20b obtains the distances between the points _hk and _Wk for the points _{h0 ,} .

As shown in FIG. 4A, the error area tends to increase in the order of healthy subjects, essential tremor, and cerebellar disorders. Therefore, using the first threshold T1 as a reference, the determination unit 20b determines that there is a high possibility that the subject E is healthy when the error area is smaller than the first threshold T1. Further, the determination unit 20b determines whether the error area is between the first threshold T1 and the second threshold T2. For example, if the error area is greater than or equal to the first threshold T1 and less than the second threshold T2 It is determined that subject E may be suffering from essential tremor. Furthermore, the determination unit 20b determines that there is a possibility of suffering from cerebellar disorder when the error area is equal to or greater than the second threshold T2, using the second threshold T2 as a reference. The first threshold T1 can be, for example, 800 mm ² and the second threshold T2 can be, for example, 1200 mm ² . Other numerical values can be set for the first threshold T1 and the second threshold T2.

Next, a method for calculating the difference between the total length of the reference line pattern BL and the total length of the handwriting pattern WL will be described. The handwriting pattern WL is divided into minute line segments dL, and the dL are integrated. The cumulative result is the total length of the handwriting pattern WL. dL may be a length that can be approximated to a straight line. The total length of the reference line pattern BL is known, and if the total length of the handwriting pattern WL is known, the ratio of the total length of the handwriting pattern WL to the total length of the reference line pattern BL, that is, the total length of the handwriting [%] can be obtained.

As shown in FIG. 4B, the handwriting total length [%] tends to increase in the order of healthy subjects, essential tremor, and cerebellar disorder. Therefore, using the first threshold S1 as a reference, the determination unit 20b determines that there is a high possibility that the subject E is healthy when the total handwriting length is smaller than the first threshold S1. Further, the determination unit 20b determines whether or not the total length of handwriting is between the first threshold value S1 and the second threshold value S2. It is determined that subject E may be suffering from essential tremor. Furthermore, the determining unit 20b determines that there is a possibility of suffering from cerebellar disorder when the handwriting total length is equal to or greater than the second threshold value S2 based on the second threshold value S2. The first threshold S1 can be, for example, 103%, and the second threshold T2 can be, for example, 105%. Other numerical values can be set for the first threshold S1 and the second threshold S2.

Note that the determination unit 20b may use both the error area and the handwriting total length [%] to perform the determination described above. At this time, for example, if it is determined that there is a possibility of essential tremor in the error area, and it is determined that there is a possibility of cerebellar disorder in the total handwriting length [%], It can be said that there is a nature.

As described in detail above, according to the present embodiment, the symptoms of the subject E can be determined simply by having the sheet 2 and the portable terminal 10 capable of imaging the sheet 2 at hand of the subject E. It can be carried out. Further, according to the present embodiment, since the handwriting pattern WL drawn by tracing the reference line pattern BL is used, the symptoms of the subject E can be determined under uniform conditions for healthy subjects. Furthermore, according to the present invention, only the imaging data 2a of the handwriting pattern WL is used for determining the symptoms of the subject E. FIG. Therefore, the symptoms of the subject E can be determined more accurately and easily.

Further, according to the present embodiment, it is possible to determine not only whether the subject E is likely to be healthy, but also whether or not the subject E is likely to be suffering from essential tremor. or whether there is a high possibility that the subject E is suffering from cerebellar disorder. In this way, the type of disease of subject E can also be determined more accurately and easily.

According to the present embodiment, the mobile terminal 10 only needs to include the imaging unit 10a for imaging the sheet 2, the display unit 10b for displaying the determination result 3, and the transmitting/receiving unit 20c for transmitting/receiving data. There is no need to install software for performing complex processing such as determination of the symptoms of subject E, or the like. Further, according to the present embodiment, determination is made only by the handwriting pattern WL drawn on the sheet 2, so that it is suitable for simple determination and it is possible to shorten the time required for determination.

Furthermore, since the imaging data 2a and the judgment result 3 transmitted and received between the mobile terminal 10 and the server computer 20 can be anonymized, it is possible to prevent leakage of personal information.

The determination results of the diagnosis support system 1 can be used by ordinary people to determine whether they are healthy or require medical attention, and can be used by doctors other than neurologists as diagnostic materials for diagnosing patients. be able to.

In the above embodiment, it is determined whether the subject E is likely to be healthy, likely to have essential tremor, or likely to have cerebellar disorder. The invention is not limited to this. It may be determined whether the subject E is likely to be healthy, or whether the subject E is highly likely to have essential tremor, or whether the subject is likely to have cerebellar disorder. You can just do it.

In the above embodiment, it is determined whether the subject E is likely to be healthy, likely to have essential tremor, or likely to have cerebellar disorder. The invention is not limited to this. Each possibility may be represented by a probability. For example, an 80% chance of being healthy and a 20% chance of suffering from essential tremor may be displayed. This probability can be obtained from statistical data such as those shown in FIGS. 4(A) and 4(B).

Symptoms that can be determined by the diagnostic support system 1 are not limited to essential tremor and cerebellar disorder. The present invention is also applicable, for example, to the determination of other involuntary movements such as myoclonus and chorea.

Moreover, although the number of layers of the convolutional neural network 21 is three in the first embodiment, the present invention is not limited to this. The number of layers of the convolutional neural network 21 may be four or more. Further, as the convolutional neural network 21, an applied convolutional neural network such as R-CNN (Regions with CNN feature), Fast R-CNN, SegNet, and U-Net may be used. Also, the learning method is not limited to the error backpropagation method, and other methods can be applied.

Further, in the first embodiment, the image itself is input to the convolutional neural network 21, but the present invention is not limited to this. The parameters of the handwriting pattern WL such as the error area and the total handwriting length [%] obtained in the second embodiment may be input.

Further, in the second embodiment, the error area and total length [%] of the handwriting pattern WL with respect to the reference line pattern BL are used as index values for determination. However, the invention is not limited to this. Another index, for example, an index value (for example, the number of intersections of the reference line pattern BL and the handwriting pattern WL) that indicates whether or not the shift of the handwriting pattern WL with respect to the reference line pattern BL is small. You can do it.

In addition, in the diagnosis support system 1 according to each of the above embodiments, the reference line pattern BL is a spiral (spiral), but the present invention is not limited to this. The reference line pattern BL may be straight, circular, square or triangular. The reference line pattern BL is desirably a figure such as a spiral or an arc, in which the direction of the drawn line changes in various directions.

In addition, the hardware configuration and software configuration of the mobile terminal 10 and the server computer 20 are examples, and can be arbitrarily changed and modified.

The main processing part of the mobile terminal 10, which is composed of the control unit 31, the main storage unit 32, the external storage unit 33, the camera 34, the touch panel 35, the wireless communication interface 36, the internal bus 30, etc., is as described above. , can be realized using a normal computer system without depending on a dedicated system. The same is true for the main processing part of the mobile terminal 10, which is composed of the control unit 41, the main storage unit 42, the external storage unit 43, the communication interface 44, the internal bus 40, and the like. For example, storing and distributing a computer program for executing the above operation in a computer-readable recording medium (flexible disk, CD-ROM, DVD-ROM, etc.), and installing the computer program in the computer. The portable terminal 10 and the server computer 20 that execute the above processes may be configured by the above. Alternatively, the computer program may be stored in a storage device of a server device on a communication network such as the Internet, and the diagnosis support system 1 may be configured by downloading the program to a normal computer system.

When the functions of the computer are shared between the OS (operating system) and the application program, or when the OS and the application program work together, only the application program portion may be stored in the recording medium or storage device.

It is also possible to superimpose a computer program on a carrier wave and distribute it via a communication network. For example, the computer program may be posted on a bulletin board system (BBS, Bulletin Board System) on a communication network and distributed over the network. Then, the computer program may be configured to be started and executed in the same manner as other application programs under the control of the OS so that the above processes can be executed.

The present invention is capable of various embodiments and modifications without departing from the broader spirit and scope of the invention. Moreover, the embodiment described above is for explaining the present invention, and does not limit the scope of the present invention. That is, the scope of the present invention is indicated by the claims rather than the embodiments. Various modifications made within the scope of the claims and within the meaning of equivalent inventions are considered to be within the scope of the present invention.

INDUSTRIAL APPLICABILITY The present invention can be used in the medical field.

1 diagnosis support system, 2 sheet, 2a, 2b, 2c imaging data, 3 determination result, 10 mobile terminal, 10a imaging unit, 10b display unit, 10c transmission/reception unit, 15 network, 20 server computer, 20a correction unit, 20b determination unit , 20c transmission/reception unit, 21 convolutional neural network, 22 parameter adjustment unit, 23 teacher data, 30 internal bus, 31 control unit, 32 main storage unit, 33 external storage unit, 34 camera, 35 touch panel, 36 wireless communication interface, 39 Program, 40 internal bus, 41 control unit, 42 main storage unit, 43 external storage unit, 44 communication interface, 49 program, BL baseline pattern, BLa, BLb image, E subject, F writing instrument, P1, P2, P3 , P4 alignment pattern, P1a, P2a, P3a, P4a, P1b, P2b, P3b, P4b image, WL handwriting pattern, WLa, WLb, WLc image

Claims (10)

an imaging unit for imaging a sheet on which a reference line pattern is printed and a handwriting pattern drawn by a subject by tracing the reference line pattern using a writing instrument;
a determination unit that determines symptoms related to the brain or nerves of the subject based on imaging data captured by the imaging unit;
with
The imaging unit is mounted on a mobile terminal,
The determination unit is mounted on a server computer that can communicate with the mobile terminal,
Diagnostic support system. The determination unit determines the symptoms of the subject by performing deep learning using imaging data for which the determination result of the symptoms of the subject is known as teacher data.
The diagnosis support system according to claim 1. The determination unit is
Determining the symptoms of the subject using a convolutional neural network trained using the teacher data, with the imaging data as input and the result of determining the symptoms of the subject as output.
The diagnosis support system according to claim 2. In the judgment result,
that the subject is likely to be healthy, that the subject is likely to be suffering from essential tremor, and that the subject is likely to be suffering from cerebellar disorder high, including at least one of
The diagnosis support system according to claim 3. The determination unit is
calculating at least one of a total amount of error between the reference line pattern and the handwriting pattern and a difference in total length between the reference line pattern and the handwriting pattern as an index value based on the imaging data;
Determining the symptom of the subject based on the calculated index value;
The diagnosis support system according to claim 1. The determination unit is
If the index value is smaller than the first threshold, it is determined that the subject is likely to be healthy,
If the index value is greater than or equal to the first threshold and less than the second threshold, determining that the subject is likely to be suffering from essential tremor,
If the index value is equal to or greater than the second threshold, it is determined that the subject is likely to be suffering from cerebellar disorder.
The diagnosis support system according to claim 5. Alignment patterns are printed on the sheet at vertices of a virtual rectangle containing the reference line pattern,
The imaging data includes an image of the alignment pattern together with the image of the handwriting pattern,
a correction unit that corrects the imaging data based on the relative positional relationship of the image of the alignment pattern in the imaging data;
The determination unit determines the symptoms of the subject based on the imaging data corrected by the correction unit.
The diagnostic support system according to any one of claims 1 to 6. A first of the alignment patterns is printed at a first density when expressed in gray scale, and a remaining second alignment pattern is printed at a density lower than the first density. It is printed with a density of 2,
The reference line pattern is printed at the second density when expressed in grayscale,
The handwriting pattern is drawn with a third density between the first density and the second density when expressed in grayscale,
The correcting unit extracts the image of the handwriting pattern based on a change in luminance in a range from the luminance of the image of the first alignment pattern to the luminance of the image of the second alignment pattern, binarize the imaging data,
The determination unit determines whether the subject is healthy based on the imaging data binarized by the correction unit.
The diagnosis support system according to claim 7. the baseline pattern is a spiral;
The diagnostic support system according to any one of claims 1 to 8. A computer that operates as a mobile terminal,
Functioning as an imaging unit for imaging a sheet on which a reference line pattern is printed and a handwriting pattern drawn by a subject by tracing the reference line pattern with a writing instrument is formed,
a server computer capable of communicating with the mobile terminal,
Based on the imaging data captured by the imaging unit, it functions as a determination unit that determines symptoms related to the brain or nerves of the subject.
program.

Images