KR102128435B1 - Artificial intelligence system for depression based on patient diary record - Google Patents

Abstract

The present invention relates a psychological analysis system recognizing the record of patient struggling against disease which includes: a preparatory step of allowing a main server to sample and store selected words, colors, and objects for mental and psychological analysis based on the big data stored in a storage device; a recording step of capturing, saving and recording a paper sheet, a word file, a recording file, a video file, or a recorded photo, which is a record of patient struggling against disease from a main server; a conversion step in which the main server divides and converts a recorded sheet or file into text and picture form so that the main server can recognize the recorded sheet or file through a wired or wireless internet network; a storage step of storing the words, colors and the number of times extracted from the converted data as stored words and colors in the main server; a comparison step of comparing the extracted words, colors, and stored words of objects with previously extracted and stored big data; a diagnosis step of marking a comparison result as a probability value and recognizing the probability value as a symptom determination value; and a prescribing step of prescribing a treatment plan based on the big data on the main server according to the diagnosis.

Description

Psychological analysis system for automatically recognizing and analyzing patient-related records {Artificial intelligence system for depression based on patient diary record}

The present invention relates to a psycho-psychological analysis system that automatically recognizes and analyzes patient-related records.

In general, suicide attempts are a worldwide problem, and attempts to recognize and predict them as machine coldsights began on May 5, 2017 under the title <Machine Learning Might Help Identify Those Most At Risk of Suicide> https://www.psychologicalscience. org/publications/observer/obsonline/machine-learning-might-help-identify-those-most-at-risk-of-suicide.html .

A paper on depression began on June 26, 2018 at https://www.mjmmed.com/article?articleID=34 under the title <Deep Learning: A New Horizon for Personalized Treatment of Depression?>.

Mental health-related applications can be found at https://www.medicalnewstoday.com/articles/320557.php .

An attempt to analyze depression in relation to color is titled <the correlation between color choices and impulsivity, anxiety and depression> http://www.ejgm.co.uk/The-correlation-between-color-choices-and-impulsivity- nanxiety-and-depression,81905,0,2.html .

However, this is just a thesis level, and no attempt has been made to use it for treatment by predicting psychoanalysis such as depression.

The present invention seeks to solve this, and the object of the present invention is to define the mental psychology of patients, such as depression, using words, colors, pictures, and specific behaviors as judgment factors, and predict mental states based on the patient's journal. The aim is to provide a psychological analysis system that automatically recognizes and analyzes patient-related records that can be analyzed and dealt with.

Another object of the present invention is to define psychological psychology based on words, specific behaviors, etc. by defining psychological psychology of patients, such as depression, using words, colors, pictures, specific behaviors, etc., and capturing relevant records written by patients. The purpose of this study is to provide a psychological analysis system that automatically recognizes and analyzes patient-related records that determine whether to deteriorate or alleviate patient treatment directions.

The present invention for this purpose, the main server is a preparatory step of sampling and storing selected words, colors, pictures, specific behaviors, etc. based on big data stored in a storage device;

A recording step of storing and recording the patient's disease journal in the form of a paper sheet, a word file, a recording file, a video file, or a picture file created directly or offline on an online platform;

A conversion step of converting the recorded sheet or file into text, color, picture, video, etc. by the user terminal or the main server so that the main server can recognize it through a wired or wireless Internet network;

A storing step of storing words, pictures, specific actions and the number of times extracted from the converted data in the main server;

A comparison step of comparing extracted recognition words, recognition colors, specific behavior recognition images, and the like and stored big data;

A diagnosis step of marking the comparison result as a probability value and diagnosing the probability value as a symptom judgment value; And

It includes a prescribing step of prescribing treatment methods based on big data from the main server according to diagnosis, and provides a psychological analysis system that recognizes and analyzes records related to patient disease.

As described above, the present invention defines psychological psychology of patients, such as depression, using words or specific behaviors or colors as judgment factors, and enables the patient to predict and analyze mental states based on a journal written by the patient.

The present invention defines the psychological psychology of patients, such as depression, using words or specific behaviors, colors, etc. as a judgment factor, and analyzes whether psychological deterioration or deterioration is based on words extracted by capturing journals written by patients and analyzing the patient's treatment direction It is available to determine.

1 is a system configuration diagram of the present invention,
2 is a flow chart of the present invention,
3 is a detailed system configuration diagram of the main server of the present invention.

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

In the present invention, as shown in the system configuration of FIG. 1 and the flowchart of FIG. 2, the main server 20 samples psychological and psychological analysis selected words, colors, pictures, and specific actions based on big data stored in the storage device 30. Prepared to store (S10);

A recording step (S20) in which the main server 20 stores and records the patient's related records in the form of a paper sheet, an online platform, or in the form of a word file, a recording file, an image file, or a picture file stored offline;

Conversion step of converting the recorded sheet or file into text, color, picture, video file, etc. in the user terminal 10 or the main server 20 so that the main server 20 can recognize it through the wired/wireless internet network 40 (S30);

A storing step (S40) of storing words, colors, pictures, specific action images, etc., extracted from the converted data, and the number of times as recognition words, recognition colors, and recognition images in the main server 20;

A comparison step (S50) of comparing the extracted recognition word, recognition color, recognition image, etc. with previously stored big data;

A diagnosis step of marking the comparison result as a probability value and diagnosing the probability value as a symptom judgment value (S60); And

In accordance with the diagnosis, a method for treatment is configured including a prescribing step (S70) for prescribing a method based on big data in the main server.

3 is a system configuration diagram for searching and collecting in the main server 20 and storing it in the storage device 30. The main server 20 shows words, pictures, images, and the number of times extracted from the journal converted in the conversion step S30. ) Is a configuration diagram to perform the storage step (S40) for storing as a storage word,

Psychopsychological analysis system for analyzing and analyzing patient logbooks of the present invention,

The main server 20 samples and stores selected words, colors, objects, specific behaviors, etc. based on big data stored in the storage device 30, psychological and psychological analysis,

Save on the main server as a patient's related paper sheet, a word file created directly or offline on an online platform, or a recorded picture, a recorded file, or a video file.

A search module (50) for retrieving recognition words, recognition colors or recognition images corresponding to depression in order that the main server divides and converts the recorded pictures or files into text and recording files, video files, and picture files;

A multiplexer 60 sequentially selecting whether there is a search word in the search module 50;

A determination module 70 for determining and storing words, colors, or objects selected by the multiplexer 60 as words, colors, and specific actions; And

It comprises a storage device (30) provided to the main server (20) to determine the degree of symptoms by dividing and storing words, colors, objects, and specific actions determined by the determination module (70) by field. .

The search module 50 includes a word search module 51 that searches for words in word files (*,txt, *.doc, *.hwp, *.html);

A voice recognition module 52 for retrieving text using voice recognition technology from recorded files (*.wma, *.pcm, *.aac, *.wma, *.wav, *.mp3) or video files;

A color recognition module 53 for recognizing colors based on Pantone colors in picture files (*.png, *.gif, *.tif, *,jpg, *.pdf); And

By separating fixed and moving images from picture files (*.png, *.gif, *.tif, *,jpg, *.pdf) or video files (*.avi, *.mp4), words of things or specific action videos It may be exemplified by including the object recognition module 54 to recognize the recognition image of.

The color recognition in the present invention is based on chromotherapy, and chromotherapy refers to diagnosis and treatment of psychology through summary colors, and refers to psychological diagnosis and treatment through color transfer. The human psychology is diagnosed and treated through a mechanism that occurs in the brain while a human reacts to color. It's not a cure for the sick, but it can help with psychological disorders, such as depression. It mainly uses red, yellow, and blue, and each color has the effect of making you feel rest, vitality, stability, energy, calmness, freedom, etc. Although not generalized in Korea, it is widely accepted as a popular therapy in Western and Japanese, and is also accepted as an alternative medicine (see A critical analysis of chromotherapy and its scientific evolution).

On the basis of this background, the present invention allows the computer to recognize colors based on the pantone color that can be recognized by the computer, and the pantone color can be exemplified by classifying small tables 1 to 5 in Table 1 below.

The words used by patients with depression in the present invention are Jeong Seon-yu, a Health Chosun intern reporter, and Kim Jin-gu, a Health Chosun reporter 2018.10.16. 16:35 ⓒ Provided by: Health Chosun,

Research has shown that words used in SNS can predict the occurrence of depression.

Researchers at the University of Pennsylvania and Stony Brook University in the United States have analyzed Facebook's records of 1175 people to see if they can predict words and depression in social media including Facebook. 114 of the participants were diagnosed with depression.

Excluding participants who recorded less than 500 words on SNS, 114 participants diagnosed with depression were compared with 5 participants without depression. The researchers analyzed the content, length, and frequency of postings on Facebook to predict depression.

As a result of the study, it was possible to predict depression three months before being diagnosed with depression. Words to predict depression included cognitive things like "sadness" as emotionally, "lonely" and "enmity" such as interpersonal, 'preconceptions about self.

Users also receive a diagnosis of depression appeared to be using a lot of words, such as first-person "I". "There is a perception that using social media is not good for mental health, but it can be an important tool for diagnosis, monitoring and treatment." The results of this study were published in the world-renowned journal Proceedings of the National Academy of Sciences.

In addition, the results of research that identified words and styles commonly used by depression patients using deep learning artificial intelligence (AI) and computer algorithms in clinical psychology (Clinical Psychological Science) in the medical journal (2018, 2, 12) are published. It's done .

The content of the presentation was that depression leaves a trace in all aspects of the patient, including “mental,” emotional, “physical,” and “language”. The research team used a software called “Languistic”Inquiry　and　Word　Count (　LIWC) to analyze the language used by patients with depression. LIWC can capture the “type” of words, “frequency” in which words are used, “vocabulary” diversity, “average” length of sentences, and “grammatical” patterns. The research team took advantage of this and analyzed the personal essays and journals written by patients with depression, to understand the characteristics in terms of content and style.

1. Description: Depression patients often write their feelings in writing. In particular, many words such as'lonely','sad'and'miserable' are used, and many adjectives and adverbs with negative meanings are used to express emotions. 　 Depression patients also used the words'I'(I),'I' (me), and'Myself' (myself). It means that the center of the mental world is tailored to yourself, and that you are not having a close relationship with those around you. The use of pronouns rather than negative words is a better clue to identify depression patients.

2. “Style”: “Depression” patients express themselves in a unique way and look at the world. As a result of analyzing the 64 online “mental health” forums with 6,400 attendees, it was found that “there are many more assertive” words. In other words, patients with “depression” often use a lot of “assertive” words, such as “completely” or “nothing”. This means that “depression” patients see the world divided into “black and white” from their own perspective. When compared to articles written by the general public, there were more than 50% more assertive expressions in the articles that appeared on the Forum of Anxiety and Depression, and 80% more in the articles that appeared on the Self-Imagination Forum.

Against this background, the present invention allows this to be performed in the main server 20, and the big data is stored in the storage device 30, and can be used for treatment based on the patient's fighting log or fighting recording, fighting images, and fighting pictures. It was completed.

Specifically, as shown in FIGS. 1 and 2, the main server 20 samples, stores, and updates selected words, colors, and objects to enable psychological and psychological analysis based on big data stored in the storage device 30. (Preparation step (S10)).

Then, on the main server 20 through the user terminal 10, the patient's fighting log (recording step (S20)), a paper sheet capture file (for example, *jpg, *.pdf, etc.) or an online platform can be created or offline Word files (*,txt, *.doc, *.hwp, *.html, etc.) created on the screen, or recorded picture files (*.png, *.gif, *.tif, *,jpg, *.pdf, * .ai, etc.) or recorded files (*.wma, *.pcm, *.aac, *.wma, *.wav, *.mp3, etc.) or video files (*.avi, *.mp4, etc.) It will be understood that it is not limited to this) (converting step (S30)) and save,

The main server 20 converts the stored photo or file into text and recording files, video files, and picture files, and in turn searches for recognition words, recognition colors, specific behavior recognition images, etc. corresponding to depression. ). The search module may be built in the main server 20 or additionally installed or may be implemented in a program form.

Subsequently, in the multiplexer 60, the search module 50 sequentially scans and selects whether there is a search word.

Subsequently, the determination module 70 judges the selected word or color, object or recognition word, recognition color, or specific behavior recognition image from the multiplexer 60 and stores it in the storage device 30 to minimize the amount of data (storing step) (S40)).

The main server 20 classifies the recognition word, the recognition color, and the recognition image determined by the determination module 70 into fields and stores them in the storage device 30 and stores the number of times together to determine the degree of depression (comparative step ( S50), diagnostic stage (S60)).

Subsequently, based on the patient's condition, the main server 20 presents a customized prescription of the patient (prescription step S70).

In the above, in the search module 50, a picture file (*.png, *.gif, *.tif, *,jpg, *.pdf, *.ai, etc.) or a video file (*.avi, *.mp4) is main On the basis of the big data previously stored in the server or the storage device 30, it is possible to recognize a recognition word or a recognition image of an object, and to recognize a recognition word and a recognition image by separating a fixed from a moving one for easy recognition. The object recognition module 54 is used. This makes it easy to recognize objects and shorten object recognition time by distinguishing background objects (eg mountains, rivers, roads) from moving objects (eg birds, trains, buses, airplanes, etc.) in video files. Because.

10; user terminal 20; main server 30; storage device 40; internet network 50; search module, 51; word search module 52; voice recognition module 53; color type module 54; object recognition module 60; multiplexer 70; judgment module

Claims (3)

delete The main server 20 samples and stores selected words, colors, pictures, specific behaviors, etc. based on the big data stored in the storage device 30, and
On the main server, save it as a paper sheet that is a record of the patient's disease or a word file written on an online platform or recorded offline, or as a recorded picture, a recorded file, or a video file.
A search module (50) for sequentially retrieving recognition words, recognition colors, or recognition images corresponding to depression based on the conversion of the stored files into text files, recording files, and picture files;
A multiplexer 60 sequentially selecting whether a search word or a color or image is present in the search module 50;
A judgment module 70 for comparing and recognizing the selected recognition word or recognition color or recognition image from the multiplexer 60 with big data previously stored in the main server; And
Characterized by including the storage device 30 provided to the main server 20 to determine the degree of symptoms by storing the number of words, colors, and recognition images, etc., determined by the determination module 70, by field, and storing the number of times. A psycho-psychological analysis system that automatically recognizes and analyzes records related to phosphorus and patient disease.
According to claim 2, Search module (50) The word search module (51) for searching for words in word files (*,txt, *.doc, *.hwp, *.html);
A voice recognition module 52 for retrieving characters using voice recognition technology from a recording file (*.wma, *.pcm, *.aac, *.wma, *.wav, *.mp3);
A color recognition module 53 for recognizing colors based on Pantone colors in picture files (*.png, *.gif, *.tif, *,jpg, *.pdf, *.ai); And
Words of things by separating fixed and moving from picture files (*.png, *.gif, *.tif, *,jpg, *.pdf, *.ai) or video files (*.avi, *.mp4) B. A psycho-psychological analysis system that recognizes and analyzes a patient's fighting log, comprising an object recognition module (54) that recognizes a specific behavior recognition image.

Images