TWI644283B - Treatment method and system for cognitive impairment - Google Patents

Abstract

The invention provides a method for treating a dynamfunction, comprising the steps of: receiving a speech signal and converting it into a speech feature function; and performing a recognition function of the speech according to the speech feature parameter to generate a recognition result and a linguistic identity; The identification result generates a first associated information, and is converted into a display information presented on a screen for the patient to view; and according to the identity of the speaker, a second associated information is downloaded from a social networking website and converted into the The display information is presented on the screen for the patient to watch.

Description

Treatment methods and systems for dysfunction

The present invention relates to a method and system for treating a dynamic disorder, and in particular to a method and system for treating a linguistic dysfunction with a related function of a linguistic recognition technology and a network community platform.

The World Health Organization (WHO) defines the age of 65 as the old man, which is generally known as the silver-haired family. At the end of 1993, China’s elderly population exceeded 7% of the total population and officially became an aging society. At present, the population of Taiwan over 65 has reached 12%. Based on this trend, the population of the elderly will reach 14% in 2017, becoming an “aged nation”, even reaching 20% in 2025 and 39.4% in 2060, becoming a “super aged nation”. .

Mild Cognitive Impairment (MCI) refers to a memory that is more severely degraded than a normal person with the same educational level and age, but there is no other symptom of obvious dementia. However, studies have pointed out that MCI may be a symptom of early dementia or a higher probability of dementia than the average person. According to the Taiwan Population Estimation Report from 2010 to 2060, the National Population Growth Estimation Report is the number of mothers and the loss of all ages. The prevalence rate of the wisdom community is estimated. The prevalence rate of Taiwan’s silver-haired dementia is as follows: 1.2% for 65-69, 2.2% for 70-74, 4.3% for 75-79, and 8.4% for 80-84. The age of 85-89 is 16.3%, and that of 90-year-old is 30.9%. The higher the age, the higher the prevalence of dementia. In addition, according to the 2011 estimates of the Economic Development Association, the number of people with mental disabilities in Taiwan will reach 800,000 in 2060. Compared with the information published by the World Health Organization (WHO), the proportion of people with dementia grows at an alarming rate, much higher than the global . And according to estimates, about 10%-15% of patients with mild dysfunction will develop into Dementia patients every year.

According to the above description, with the aging trend, if the population of mild dysfunction (MCI) does not take appropriate measures to help slow down the recession, it is estimated that 10%-15% of them will become dementia patients.

Furthermore, dementia is a brain-wide cognitive function syndrome characterized by memory loss. Other possible symptoms include abstract thinking, judgment, orientation, language and other cerebral cortex, and may also incorporate personality. Changes, disturbances, psychotic symptoms, social dysfunction, and decreased self-care ability, so it affects patients' memory, concentration, language skills, and problem-solving skills. With the trend of aging, the care of the elderly with dementia is becoming more and more serious. The cost of care for the dementia population is very large, and many problems arising from it will be delayed by the national economy and finance if not properly handled.

It is an object of the present invention to provide a method and system for treating a dysfunctional disorder. The use of Information and Communication Technology (ICT) technology, combined with the association between the speaker identification technology and the online community platform, has become a treatment and system for mild dysfunction disorders to delay patients with mild dysfunction ( Or the degree of decline in the knowledge of the silver babies with mild dynamfunction.

The present invention is a dysfunctional treatment system for treating a patient, the treatment system comprising: a handheld device having a processing circuit for receiving a speech signal of a speaker and converting to a speech feature function, and a The screen displays a display information on the screen for the patient to view; a server receives the voice feature parameter and performs a speaker recognition action to generate a recognition result and the speaker identity, and correspondingly associates a first associated information Converting to the display information; and a social networking website receiving the identity of the speaker provided by the server and generating a second associated information to the server, so that the server converts the second associated information into the display information.

The invention relates to a dysfunction treatment method for treating a patient, the treatment method comprising the steps of: receiving a speech signal and converting into a speech feature function; and performing a speech recognition function according to the speech feature parameter, generating Identifying the result and the identity of the speaker; generating a first associated information according to the identification result, and converting the display information to a display for viewing by the patient; and downloading from a social networking site according to the identity of the speaker A second associated information is converted and the display information is presented on the screen for the patient to view.

In order to better understand the above and other aspects of the present invention, the following detailed description of the embodiments and the accompanying drawings

110‧‧‧ tablet

112‧‧‧ screen

114‧‧‧Processing Circuit

120‧‧‧Server

122‧‧‧ Identification unit

124‧‧‧Information acquisition unit

128‧‧‧Database

130‧‧‧Community website

S210~S216‧‧‧Step process

FIG. 1 is a schematic diagram of a treatment system for mild dysfunction according to the present invention.

Figure 2 is a flow chart showing the treatment method of the dysfunction of the present invention.

Because the most prominent feature of patients with Mild Cognitive Impairment (MCI) is a significant decline in memory, it can seriously affect interpersonal relationships.

Reminiscence Therapy, also known as the Remembrance Therapy, is defined as a mental process by recalling past events of special significance to the individual, meaning that the patient recalls, discusses, and shares the life experiences of the past in an organized way. Lieutenant reorganized past events and experiences. Recalling past years and achievements can enhance patients' self-esteem, self-confidence, personal dignity and relief of depression, and also enhance patient communication with others and affirm the meaning of life. By selecting topics of interest to the patient and using "guides" such as photos, old songs, patterns, movies, souvenirs, etc., topics such as marriage customs, childhood memories, birthday parties, classmates, competition experiences, etc. are introduced. Encourage and guide the patient's memories.

The omnipotent treatment method and system of the present invention utilizes the correlation function between the speaker recognition technology and the online community platform to improve the memory of MCI patients, thereby improving their interpersonal relationship and slowing the speed of dementia. The following is a detailed description: Please refer to FIG. 1 , which is a schematic diagram of a treatment system for mild dysfunction according to the present invention. The treatment system of the present invention includes a tablet 110, a server 120, and a social networking site 130. Among them, the MCI patient carries the tablet 110 with him, and the server 120 is connected to the social networking site 130.

The processing circuit 114 in the tablet 110 receives and processes the speech signal Si of the visitor (i.e., the speaker) to generate the speech feature parameter Sp to the server 120. Of course, the tablet 110 can be replaced with other hand-held devices that are similar in function.

The identification unit 122 in the server 120 receives the speech feature parameter Sp and generates a recognition result Sr. The data acquisition unit 124 generates the speaker identity V (that is, the identity of the visitor) based on the recognition result Sr.

Basically, the database 128 inside the server 120 can generate the first associated information I1 according to the identification result Sr or the speaker identity V. The first associated information I1 includes the identity and background information of the visitor, and the server 120 converts the first associated information I1 into the display information D and transmits it to the tablet 110. The tablet 110 can display the identity and background information of the visitor on the screen 112, so that the MCI patient can quickly remember the identity of the visitor.

In addition, the data extraction unit 124 in the server 120 transmits the speaker identity V to the community website 130, and the community website 130 downloads the second association information I2. Similarly, the server 120 converts the second associated information I2 into the display information D. Transfer to tablet 110. Basically, the second associated information I2 is historical data stored by the visitor in the social networking website 130 regarding MCI patients and visitors. For example, the second associated information I2 includes a photo or movie of an MCI patient meeting with a visitor, traveling, or participating in an activity. The tablet 110 can display a photo or video of the interaction between the MCI patient and the visitor on the screen 112.

Since nostalgic therapy can help improve the memory of MCI patients, the treatment system of the present invention can recognize the speaker's identity V based on the visitor's speech signal Si. Since the database 128 of the server 120 has established its own first associated information I1 for the visitor, the information retains the most familiar memories between the MCI patient and the visitor (friends and relatives), so the first related information I1 is converted. In order to display the data D and display it on the screen 112 of the tablet 110 for the patient to watch and recall, the strange feeling between the patient and the visitor can be quickly disappeared.

In addition to the first associated information I1, the data capture unit 124 of the server 120 automatically retrieves the second associated information from the cloud social network 130. After displaying the first associated information I1, The content of the second related information I2 is displayed on the screen 112 of the tablet 110 through the display information D. Through the use of this information, nostalgic therapy can play a role in improving the interpersonal relationship of MCI patients, and can slow down the speed of dementia.

According to an embodiment of the present invention, the processing circuit 114 of the tablet computer 110 can receive the voice signal Si of the visitor, and convert the voice signal Si into the voice feature parameter Sp by using Mel-Frequency Cepstrum Coefficients (MFCCs). And upload to the server 120 executor Identify the action. Of course, the present invention is not limited to the use of MFCCs to generate the speech feature parameter Sp, and other methods may be used to generate the speech feature parameter Sp.

The identification unit 122 and the data capture unit 124 in the server 120 can be a hardware circuit (such as a central processing unit CPU of the server 120) or a software program executed by the central processing unit. Basically, the present invention utilizes a Gaussian Mixture Model (GMM) and a Universal Background Model (UBM) to build a Speaker Model. Of course, the present invention is not limited to this. For example, the GMM/UBM can also be used to build a speaker model, and then included in a Support Vector Machine (SVM).

The identification unit 122 uses a Gaussian mixture model (GMM) to establish a set of text-independent speaker recognition system. The GMM speaker model is based on a number of numbers to form a string of corpus as a basic training unit. Through the LBG (Linde, Buzo and Gray) algorithm and EM (Expectation Maximization) algorithm to train each speaker's GMM, each GMM tentatively set 8 to 256 Gaussian mixed numbers (Mixtures), will be based on actual performance indicators Adjust the number of blends.

For example, when all the speaker GMMs are created, there are N speaker GMMs (ie, λ ₁ , λ ₂ , . . . , λ _N ). After the recognition unit 122 receives the speech feature parameter Sp, the recognition result Sr is generated from the N speaker GMMs according to the Maximum-Likelihood Classification.

That is, the identification unit 122 already has a set of GMMs of the N linguists, (λ ₁ , λ ₂ , ..., λ _N ) the speaker model. The main purpose of the identification unit 122 is to find a speech feature parameter Sp={c ₁ , c ₂ , . . . , c _T } corresponding to the input, and a speaker model having the greatest probability of posterior probability. The minimum error Bayes' Decision Rule can be used to obtain the identification result Sr, which is the most likely speaker (or visitor). Therefore, the database 128 of the server 120 can generate the first associated information I1, and the server 120 converts the first associated information I1 into the display information D and transmits it to the tablet 110.

In addition, many community websites today provide developers with an Application Programming Interface (API) to access services, including social networking sites such as Windows Live, Google, Facebook, Plurk, and Twitter. Therefore, the data extraction unit 124 obtains the second association information I2 according to the API provided by the community website, and the server 120 converts the second association information I2 into the display information D and transmits the information to the tablet 110.

For example, the Facebook Graph API is a low-level Hypertext Transfer Protocol (HTTP)-based API for accessing and writing Facebook information. The book program, its functions include querying the required information, posting posts, managing advertisements, uploading photos, and so on. Since the Facebook Graph API is HTTP-based, it is compatible with all programming languages that support the HTTP library, such as eURL, urllib, etc. Therefore, the data capturing unit 124 of the present invention The Facebook Graph API can be used, and the second affiliate information I2 is downloaded from the social networking website 130, and the content is the photo related to the visitor and the MCI patient.

Please refer to FIG. 2, which is a flow chart of a treatment method for the dysfunction of the present invention. First, a speech signal Si is received and converted into a speech feature parameter Sp (step S210). Then, the speaker recognition operation is performed according to the voice feature parameter, and a recognition result Sr and a speaker identity V are generated (step S212).

Furthermore, a first associated information I1 is generated according to the identification result, and converted into a display information is presented on the screen (step S214) for the patient to view. In addition, by using the speaker's identity V, a second association information I2 is downloaded from a cloud website, and converted into the display information and presented on the screen (step S216) for the patient to view.

As can be seen from the above description, the treatment system and method of the present invention can recognize their identity through the voice of a speaker (friends or relatives), and can be linked to the self-built database 128 to immediately display the background information of the visitor (friends or relatives). And the historical social data associated with the two parties may be retrieved by the cloud social network 130, including static photos and dynamic multimedia information, such as photos and videos of parties, outings, and participating activities. Through these methods, MCI patients can immediately know the identity of the speaker (friends or relatives/visitors), and then the dice can be opened, thus avoiding the embarrassing scene and enhancing interpersonal relationships. In addition, through the related information provided by the community website 130, MCI patients can immediately return to the good memories of both parties to achieve the nostalgic therapy effect, which stimulates the memory ability of MCI patients, thus delaying the extent of knowledge degradation.

In conclusion, the present invention has been disclosed in the above embodiments, but it is not intended to limit the present invention. Those of ordinary skill in the art to which the present invention pertains, Various changes and modifications can be made without departing from the spirit and scope of the invention. Therefore, the scope of the invention is defined by the scope of the appended claims.

Claims (5)

A dysfunction treatment system for treating a patient, the treatment system comprising: a handheld device having a processing circuit for receiving a speech signal of a speaker and converting to a speech feature function, and a screen to display The information is presented on the screen for the patient to view; a server receives the voice feature parameter and performs a speaker recognition action to generate a recognition result and the speaker identity, and correspondingly converts a first associated information into the display Information, wherein the first associated information includes the identity or background information of the language; and a social networking website receives the identity of the speaker provided by the server and generates a second associated information to the server, such that the server Converting the second associated information into the display information, wherein the second associated information includes a photo or a movie of interaction between the patient and the speaker. The processing system of claim 1, wherein the server comprises: an identification unit, receiving the voice feature parameter and performing the speaker recognition action and generating the identification result; and a data acquisition unit, according to the identification As a result, the identity of the speaker is generated to the community website; and a database generates the first associated information based on the identification result. The treatment system of claim 2, wherein the identification unit receives the speech feature parameter and generates the recognition result from a plurality of speaker Gaussian mixture models according to a maximum similarity grouping method. The treatment system of claim 2, wherein the data acquisition unit obtains the second association information according to an application interface. The treatment system of claim 1, wherein the handheld device is a tablet computer, and the processing circuit generates the voice signal into the voice feature parameter by using a Mel Cepstrum coefficient.