JP6699825B2 - Diagnostic device, diagnostic device control method, and diagnostic program - Google Patents

Description

  The present invention relates to a diagnostic device and the like for diagnosing a disease in which symptoms appear in language.

  It is known that dementia correlates with verbal ability because the number of words that can be used tends to decrease in patients with dementia as the symptoms progress. Here, the measurement of language ability has attracted attention as a method for early detection of dementia because it has no physical invasion.

  Further, as a method for early detection of dementia, the Hasegawa-type simplified intelligence evaluation scale (HDS-R) and Me-CDT are known, and using these methods, elderly people suspected of dementia and normal A study that extracts the tendency and characteristics of speech with the elderly has also been reported (Non-Patent Document 1).

Akiko Shikata, Mai Miyabe, Yasuha Noda, Aya Kinoshita, Eiji Aramaki, "A Qualitative Study of Speech Text for People with Mild Dementia: For Rapid Non-Invasive Screening of Dementia," Research Report of Information Processing Society of Japan, Vol. 2015-UBI-47 No. 4 Vol. 2015-ASD-2 No. 4, July 27, 2015

  Generally speaking, linguistic ability is roughly divided into syntactic ability and vocabulary ability, but it is known that vocabulary ability is useful for dementia. As a method of measuring vocabulary ability, a method of measuring the amount of vocabulary that can be understood or used and a method of measuring the number of propositions used in authentication psychology are known.

  However, in the method of measuring vocabulary, since the unit of the word differs depending on the language, there is a problem that there is ambiguity in the unit of the word, and there is a problem that the algorithm needs to be changed according to the language. ..

  Further, the method of measuring the number of propositions has a problem that manual work by an expert is required. Although a method called CPIDR that automatically analyzes the number of propositions by counting part-of-speech has been proposed, this method is for English, and it is difficult to apply it to another language as it is.

  Further, in Non-Patent Document 1 described above, only a feature of speech between a dementia patient and a healthy person is extracted using a known evaluation method, and there is no disclosure about a new method for determining dementia.

  An object of the present invention is to provide a technique capable of easily and accurately determining whether or not a person has a suspicion of a disease in which a symptom appears in a language without depending on the language.

A diagnostic device according to one aspect of the present invention is a diagnostic device for diagnosing a disease in which symptoms appear in language,
An acquisition unit that acquires text data indicating the utterance content of the person or the document transcribed by the person,
A compression unit that reversibly compresses the text data acquired by the acquisition unit by a data compression method in which the compression rate increases as the number of repetitions of the same word increases ;
A compression rate calculation section for calculating the compression rate of the text data losslessly compressed by the compression section;
And a determination unit that determines whether or not the person has a suspicion of the disease based on the compression ratio calculated by the compression ratio calculation unit.

  It is known that when analyzing the utterance content of a patient who has a suspicion of a disease in which symptoms such as dementia appear in the language, the same word or the same phrase tends to be repeated and redundancy is high. In the lossless compression algorithm, since the original data is compressed by extracting a redundant phrase and assigning a short code to the phrase, the compression rate becomes higher as the original data contains more redundant phrases. Therefore, comparing the compression ratios when lossless compression was performed on text data indicating the utterance contents of a person suspected of having a symptom in language and a person not having a suspicion, the person having a suspicion of a disease having symptom in language was compared. The present inventor has found that is higher than that of a person who has no doubt.

  In this aspect, it is determined whether or not the person has a suspicion of a disease in which a symptom appears in the language, based on the compression rate when the text data indicating the utterance content of the person is losslessly compressed. Therefore, it is possible to determine whether or not a person has a suspicion of a disease in which a symptom appears in a language by a common algorithm without depending on the language. As a result, it is possible to easily and accurately determine a disease in which symptoms appear in language.

In the above aspect,
A sound pickup unit that picks up the voice uttered by the person,
A voice recognition unit that converts the collected voice into the text data and outputs the voice data to the acquisition unit by recognizing the collected voice by voice may be further provided.

  According to this aspect, since the text data is acquired by recognizing the voice uttered by the person, the text data can be acquired without imposing an excessive burden on the person.

In the above aspect,
The acquisition unit may acquire the text data by text-recognizing a document in which the utterance content of the person is described or a document transcribed by the person.

  According to this aspect, since the text data is acquired by recognizing the text in which the utterance content of the person is described or the document transcribed by the person, the text data can be acquired by a simple process.

  Further, in the above aspect, the determination unit may determine that the person has a suspicion of the disease when the compression rate is higher than a predetermined threshold value.

  According to this aspect, if the compression rate is higher than the predetermined threshold value, it is determined that the person has the suspicion of the disease, and thus the disease in which the symptom appears in the language can be determined by a simple algorithm.

  Further, in the above aspect, the determination unit may determine that the degree of the disease is higher as the compression rate is higher.

  According to this aspect, it is possible to present not only the presence or absence of the disease but also to what extent the person has the suspicion of the disease.

  Further, in the above aspect, the disease may include dementia, brain disease, and mental disorder.

  According to this aspect, it is possible to accurately determine dementia, brain diseases, and mental disorders that have a high correlation with language ability.

  According to the present invention, it is possible to easily and accurately determine whether or not a person has a suspicion of a disease in which a symptom appears in language without depending on language.

It is a block diagram which shows the whole structure of the diagnostic device 1 which concerns on Embodiment 1 of this invention. It is a figure which shows an example of the compression process by a compression part. 5 is a flowchart showing an example of processing of the diagnostic device according to the first embodiment of the present invention. 6 is a table summarizing language indexes used in the experiment of the present embodiment. 6 is a table summarizing the experimental results in the present embodiment. It is a block diagram which shows the whole structure of the diagnostic device which concerns on Embodiment 2 of this invention. 9 is a flowchart showing an example of processing of the diagnostic device according to the second embodiment of the present invention. It is a figure which shows the whole structure of the diagnostic device which concerns on Embodiment 3 of this invention. It is a flowchart which shows an example of a process of the diagnostic device which concerns on Embodiment 3 of this invention. It is a block diagram which shows the whole structure of the diagnostic device which concerns on Embodiment 4 of this invention. It is a flow chart which shows an example of processing of a diagnostic device concerning Embodiment 4 of the present invention.

(Embodiment 1)
1 is a block diagram showing the overall configuration of a diagnostic device 1 according to Embodiment 1 of the present invention. The diagnosis device 1 is a device for diagnosing a suspicion of a disease in which symptoms appear in language. In the following first to fourth embodiments, dementia will be described as an example of a disease in which a symptom appears in language, but the present invention is not limited to this. For example, the present invention is applicable to aphasia and developmental disorders (Asperger syndrome, learning disorder, hyperactivity disorder, etc.). That is, in the present embodiment, the disease in which language is manifested corresponds to a disease in which language is impaired due to some factor, and in addition to dementia, brain disease or mental disorder (eg, depression) in which language is manifested. Is included.

  Here, the dementia includes Alzheimer's dementia, cerebrovascular dementia, Lewy body dementia, frontotemporal dementia and the like. Further, in the present embodiment, the dementia also includes MCI (Mild Cognitive Impairment).

  In FIG. 1, the diagnostic device 1 includes a mobile terminal 100, a voice recognition server 200, and a diagnostic server 300. The mobile terminal 100 is composed of a portable information processing device including a touch panel 102, such as a smartphone or a tablet terminal. However, this is an example, and a mobile phone that does not include the touch panel 102 may be adopted as the mobile terminal 100.

  The voice recognition server 200 and the diagnosis server 300 are each configured by a computer having a communication function. The mobile terminal 100, the voice recognition server 200, and the diagnosis server 300 are communicably connected to each other via the network NT. As the network NT, a public communication network including a mobile phone communication network and an internet communication network can be adopted. The mobile terminal 100, the voice recognition server 200, and the diagnostic server 300 send and receive various data using a communication protocol such as TCP/IP.

  The mobile terminal 100 is a device that collects a voice of a person who is a diagnosis target of dementia and presents the diagnosis result to the person, and includes a sound collecting unit 101, a touch panel 102, a control unit 103, and a communication unit 104. .

  The sound pickup unit 101 includes, for example, a microphone that picks up a voice uttered by a person and converts the voice into a voice signal, a signal processing circuit that performs predetermined signal processing on the voice signal converted by the microphone, and the like. Here, the signal processing includes preprocessing such as removing noise included in the audio signal, processing of converting an analog audio signal into digital audio data, and the like.

  Under the control of the control unit 103, the touch panel 102 displays the diagnosis result transmitted from the diagnosis server 300 and displays a screen including a message prompting the person to be diagnosed to speak.

  The control unit 103 includes a CPU, a ROM, a RAM, and the like, and controls the entire mobile terminal 100. In the present embodiment, the control unit 103 transmits the voice data output from the sound collection unit 101 to the voice recognition server 200 using the communication unit 104 and the diagnosis result transmitted from the diagnosis server 300 on the touch panel 102. Perform the processing to be displayed on.

  The communication unit 104 is composed of a communication device for connecting the mobile terminal 100 to the network NT. In the present embodiment, the communication unit 104 transmits voice data to the voice recognition server 200 and receives the diagnosis result transmitted from the diagnosis server 300 under the control of the control unit 103.

  The voice recognition server 200 manages a process of converting voice data transmitted from the mobile terminal 100 into text data, and includes a voice recognition unit 201 and a communication unit 202. The voice recognition unit 201 performs voice recognition processing on the voice data transmitted from the mobile terminal 100 and converts it into text data. Here, as the voice recognition process, for example, if a known process of performing voice recognition by using an acoustic model for accumulating sound waveform data and a language model for accumulating information on words and their arrangement is adopted. Good.

  The communication unit 202 is configured by a communication device that connects the voice recognition server 200 to the network NT, receives voice data transmitted from the mobile terminal 100, and transmits the text data converted by the voice recognition unit 201 to the diagnostic server 300. Send.

  The diagnosis server 300 is a device that diagnoses whether a person has a suspicion of dementia, and includes a compression unit 301, a compression ratio calculation unit 302, a determination unit 303, and a communication unit 304.

  The compression unit 301 losslessly compresses the text data transmitted from the voice recognition server 200. Here, the compression unit 301 may use any compression method as long as it is a known lossless compression method such as ZIP, LZH, TRGZ, CAB.

  FIG. 2 is a diagram showing an example of compression processing by the compression unit 301. Generally, in lossless compression, data is compressed through two steps of dictionary method and code assignment. In the dictionary method, words that appear repeatedly in the text data are encoded in order to reduce the redundancy of the text data to be processed.

  The text data TX shown in FIG. 2 is a transcription of the contents uttered by a dementia patient. In the example of FIG. 2, in the text data TX, a uniquely identifiable code is assigned to repeatedly appearing words such as “Ah,” “Bike,” and “Toka.” As a result, the text data TX is converted into the text data TX1 as shown in the middle part of FIG.

  Next, a code allocation step is performed on the text data TX1. In the code assignment step, for example, the Huffman method is adopted, and a shorter code is assigned to a word having a higher appearance frequency. As a result, the text data TX2 is finally obtained. In this example, 65-character 130-byte text data TX is finally compressed into 43-character 86-byte text data TX2. As a result, the compression rate is 86/130=66%.

  As shown in the text data TX of FIG. 2, since a person with a suspicion of dementia has a high tendency to repeatedly use the same word, the compression ratio of the text data TX indicating the utterance content of the person with a suspicion of dementia is , Higher than the compression rate for people without dementia. Therefore, it can be determined from the compression rate whether the person to be diagnosed has a suspicion of dementia.

  This tendency appears in any compression method as long as it is a lossless compression method using the dictionary method and code assignment, and thus the present embodiment can apply the existing lossless compression method including the dictionary method and code assignment. In addition, in such an existing reversible compression method, a higher compression rate can be obtained as the number of repetitions of the same word increases, regardless of the data content to be processed. Therefore, in the present embodiment, it is not necessary to change the algorithm according to the language used by the speaker.

  In the present embodiment, the lossless compression method is adopted because the existing lossy compression method has an algorithm constructed mainly for data other than text data such as image data. Because it does not fit in.

  Referring back to FIG. The compression rate calculation unit 302 calculates the compression rate of the text data compressed by the compression unit 301. Here, the compression rate calculation unit 302 may calculate the compression rate by obtaining the ratio of the data amount of the compressed text data TX2 to the data amount of the uncompressed text data TX.

  If the compression rate calculated by the compression rate calculation section 302 is larger than the predetermined threshold value, the determination section 303 determines that the person to be diagnosed is suspected of having dementia, and if the compression rate is not higher than the predetermined threshold value. , It is determined that the person to be diagnosed has no suspicion of dementia. Here, the determination unit 303 may calculate the degree of dementia such that the value increases as the compression rate increases. For example, the determination unit 303 may calculate the dementia degree “0” indicating that there is no suspicion of dementia as the degree of dementia when the compression rate is equal to or less than the threshold value. Further, when the compression rate is larger than the threshold value, the determination unit 303 calculates the degree of dementia step by step, such as dementia degree “1” and dementia degree “2” as the compression rate increases. You may. Then, the determination unit 303 generates data indicating the determination result as the diagnosis result. Here, as the threshold value, for example, a value obtained by conducting an experiment on a large number of people, and a value that can be suspected of dementia when the value increases more than this may be adopted. ..

  The communication unit 304 includes a communication device for connecting the diagnosis server 300 to the network, receives the text data transmitted from the voice recognition server 200, and transmits the diagnosis result generated by the determination unit 303 to the mobile terminal 100. Send. Upon receiving this, the mobile terminal 100 displays the diagnosis result on the touch panel 102.

  Note that, in FIG. 1, the compression unit 301 to the determination unit 303 are realized by, for example, the CPU executing a program. Further, in the present embodiment, the communication unit 304 corresponds to the obtaining unit in the claims.

  FIG. 3 is a flowchart showing an example of processing of the diagnostic device 1 according to the first embodiment of the present invention. First. The mobile terminal 100 picks up the voice of a person to be diagnosed (hereinafter referred to as “diagnosis subject”) and acquires voice data (S101). Here, the diagnosis target person may speak, for example, according to a speech guidance message for diagnosing the suspicion of dementia, which is displayed on the touch panel 102 of the mobile terminal 100. Alternatively, the person to be diagnosed may speak through a dialogue with a doctor or a nurse (hereinafter referred to as “doctor etc.”).

  Next, the mobile terminal 100 transmits the voice data uttered by the diagnosis target person to the voice recognition server 200 (S102).

  Next, the voice recognition server 200 receives voice data (S201). Next, the voice recognition server 200 performs voice recognition on the received voice data and converts it into text data (S202). Next, the voice recognition server 200 transmits the converted text data to the diagnosis server 300 (S203).

  Next, the diagnostic server 300 receives the text data transmitted from the voice recognition server 200 (S301). Next, the diagnostic server 300 compresses the received text data (S302) and calculates the compression rate (S303).

  Next, the diagnosis server 300 determines whether the compression rate is larger than the threshold value X, and if the compression rate is larger (YES in S304), the diagnosis target person determines that the dementia is suspected (S305), and the compression rate. If is not larger than the threshold value X (NO in S304), the diagnosis target person determines that there is no suspicion of dementia (S306). Next, the diagnostic server 300 transmits the diagnostic result to the mobile terminal 100 (S307).

  Next, the mobile terminal 100 receives the diagnosis result (S103) and displays the received diagnosis result on the touch panel 102 (S104). As a result, the diagnosis result of dementia is presented to the person to be diagnosed, the doctor, and the like. In this case, the mobile terminal 100 may display a message including the degree of dementia and the description of the degree of dementia on the touch panel 102, for example. As an explanation of the degree of dementia, for example, if the degree of dementia is “0”, a message such as “normal” can be adopted, and if the degree of dementia is “1”, “the suspicion of dementia is suspected”. If there is a degree of dementia of "2", a message such as "I have a high suspicion of dementia" can be adopted.

  Next, an experiment conducted to confirm the effect of this embodiment will be described. Conventionally, various language indexes have been used as indexes for measuring the language ability of dementia patients. Therefore, in this experiment, the existing language indexes were compared with respect to how accurately the language indexes of the present embodiment, that is, the language indexes using the compression rate, can accurately diagnose dementia. In this experiment, 8 subjects were known to have a suspicion of dementia (AD; Alzheimer's disease), and 9 subjects were known to have no suspicion of dementia (nonAD). , An experiment was conducted. In addition, in this experiment, the subject was asked a question and spoken for about 13 to 17 minutes.

  FIG. 4 is a table summarizing the language indexes used in the experiment of this embodiment. In FIG. 4, “TOKEN” to “AWU” are indexes conventionally used as language indexes, and “TCR” is a language index according to the present embodiment.

  Specifically, “TOKEN” indicates the number of words spoken by the subject. “TYPE” indicates the number of types of words included in the utterance of the subject. “TTR” indicates the ratio of “TYPE” to “TOKEN”. The lower the “TTR”, the more the same word is repeated, and the higher the suspicion of dementia.

  Other details are omitted, but "TPS" (Token Per Second) is a speech rate, "LEL" (Lexical Education Level) is a difficulty level of vocabulary education, and "ADD" (Average Dependency Distance) is an average. It is a dependency distance, and “AWU” (Average Word User) is the average number of vocabulary users.

“TCR” indicates a text compression ratio (Text Compressibility Ratio), and is represented here by subscripts such as ZIP and LZH according to the compression method used. In this experiment, since ZIP, LZH, TRGZ, and CAB were adopted as the compression method, the TCRs corresponding to them are represented as TCR _ZIP , TCR _LZH , TCR _TRGZ , and TCR _CAB .

  FIG. 5 is a table summarizing the experimental results in the present embodiment. In Table 51 of FIG. 5, the first column from the left shows the language index used, the second column AD represents a subject suspected of dementia, and the third column “nonAD” indicates suspected dementia. Not showing healthy subjects.

  As shown in Table 51, the “AD” test subjects were “17.40” and “1225” as compared with the “13.27” and “814” test subjects of both “TIME” and “TOKEN” of “nonAD”. A large value is obtained. Moreover, "TTR" is as low as "0.249" as compared with "0.313" of the subject of "nonAD", and it can be seen that the number of "TYPE" is small relative to the number of "TOKEN". In addition, in Table 51, the numerical value in parentheses shows dispersion.

  In Table 51, the p-value (p value) in the fourth column is a statistical index indicating the possibility that a difference between the groups “AD” and “nonAD” may occur by chance. For example, a p-value of 0.01 (p=0.01) means that the corresponding result will occur by chance only once in 100 times. That is, the smaller the p value is, the more significant the difference between the groups is, and in this experiment, the subjects “AD” and “nonAD” can be accurately discriminated. Generally, differences between groups are interpreted as significant at p-values below 0.05.

  The table 52 in FIG. 5 is a table summarizing the details of the subjects. The first column shows the subjects with “AD” and the second column shows the subjects with “nonAD”. In Table 52, the subjects of “AD” are 1 male and 7 female, the average age is “77.2 years”, and the average MMSE score is “17.0”. The MMSE score refers to the subject's score in the dementia diagnostic test. If the score exceeds 21, it is diagnosed as "nonAD", and if it is 21 or less, the diagnosis is "AD".

  In Table 52, the subjects of “nonAD” were 4 males and 5 females, the average age was “76.6 years old”, and the average MMSE score was “25.1”. It was

As shown in Table 51, TCR _ZIP , TCR _LZH , TCR _TRGZ , and TCR _CAB have p values of "0.054", "0.015", "0.060", and "0.029", respectively. , TCR _Targz , all p values are less than 0.05, and it can be seen that the subjects “AD” and “nonAD” can be accurately discriminated. Further, although _T-CRARGZ has a p-value of more than 0.05, the excess amount is small, and it can be said that “AD” and “nonAD” can be discriminated almost accurately.

  Moreover, the p-value of “TTR” which has been conventionally used as a language index useful for diagnosing dementia is “0.02”, which shows a significant result. However, it is known that TTR largely depends on the sample size, and when the sample size is small, it becomes difficult to obtain a high p-value as in this experiment.

  As described above, since the dementia is diagnosed by using the compression rate, the diagnosis device 1 can accurately diagnose the presence or absence of dementia without depending on the language. Further, since the diagnostic device 1 is a method that does not depend on the linguistic meaning content of the utterance, even if the number of samples is small, the p-value becomes large like the TTR and the diagnostic accuracy does not deteriorate.

(Embodiment 2)
The diagnostic device of the second embodiment is characterized in that the diagnostic device is configured by a local computer without going through a network. FIG. 6 is a block diagram showing the overall configuration of diagnostic device 1A according to the second embodiment of the present invention. In the present embodiment, the same components as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted.

  The diagnostic device 1A includes a sound collection unit 410, a processing unit 420, and a display unit 412. The sound pickup unit 410 is composed of, for example, an IC recorder externally connected to a computer constituting the diagnostic device 1A, or a microphone incorporated in this computer, and picks up the voice of a person.

  The processing unit 420 and the display unit 412 are configured by a personal computer (PC) included in the diagnostic device 1A. This PC may be either a desktop PC or a notebook PC.

  The processing unit 420 includes an acquisition unit 421, a voice recognition unit 422, a compression unit 423, a compression rate calculation unit 424, and a determination unit 425. These blocks are realized by the CPU executing a diagnostic program for causing the computer to function as the diagnostic device 1A.

  The acquisition unit 421 acquires the sound data collected by the sound collection unit 410. The voice recognition unit 422, the compression unit 423, the compression rate calculation unit 424, and the determination unit 425 have the same functions as the voice recognition unit 201, the compression unit 301, the compression rate calculation unit 302, and the determination unit 303 in FIG. , Detailed explanation is omitted.

  The display unit 412 is composed of a display device such as a liquid crystal display or an organic EL display, and displays a screen having the same contents as the touch panel 102 shown in FIG.

  FIG. 7 is a flowchart showing an example of processing of the diagnostic device 1A according to the second embodiment of the present invention. First, the sound collection unit 410 collects the voice of the person to be diagnosed and acquires voice data (S801). Next, the voice recognition unit 422 converts the acquired voice data into voice data by performing voice recognition (S802).

  The processes of S803 to S807 are the same as S302 to S306 of FIG. 3, respectively. In step S808, the determination unit 425 causes the display unit 412 to display the diagnosis result.

  As described above, according to the diagnostic device 1A according to the second embodiment, since the diagnostic device 1A is configured by the local computer, the diagnostic result can be promptly obtained without depending on the communication traffic of the network.

(Embodiment 3)
The diagnostic device of the third embodiment is characterized in that the sound collecting unit is configured by an IC recorder in the diagnostic device 1 of the first embodiment. FIG. 8: is a figure which shows the whole structure of the diagnostic apparatus 1B which concerns on Embodiment 3 of this invention. In this embodiment, the same components as those in the first and second embodiments are designated by the same reference numerals and the description thereof will be omitted.

  The diagnostic device 1B differs from the diagnostic device 1 in that a PC (personal computer) 920 is used in place of the mobile terminal 100, and an IC recorder 940 is used in place of the sound pickup unit 101. .

  The PC 920 is a desktop PC or a notebook PC, and includes a display unit 921, a communication unit 922, and a USB interface 923.

  The display unit 921 includes a display device such as a liquid crystal display or an organic EL display, and displays a screen having the same contents as the touch panel 102 shown in FIG.

  The communication unit 922 includes a communication device for connecting the PC 920 to the network NT. The USB interface 923 acquires audio data from the USB memory 930. The IC recorder 940 acquires voice data by collecting the voice uttered by the diagnosis target person and records it in the USB memory 930, as in the sound pickup unit 101 shown in FIG. 1.

  FIG. 9 is a flowchart showing an example of processing of the diagnostic device 1B according to the third embodiment of the present invention. First, in S901, the PC 920 reads the voice data of the diagnosis target person picked up by the IC recorder 940 from the USB memory 930 and acquires the voice data. Thereafter, in FIG. 9, the same processing as in FIG. 3 is performed. That is, in S902 to S904 of FIG. 9, the same processing as S102 to S104 of FIG. 3 is performed, in S911 to S913 of FIG. 9, the same processing as S201 to S203 of FIG. 3 is performed, and 921 to S927 of FIG. Then, the same processing as S301 to S307 of FIG. 3 is performed.

  As described above, according to the diagnostic device 1B of the third embodiment, since the IC recorder 940 that is externally connected is used, it is possible to place the IC recorder 940 in the mouth of the person to be diagnosed, and The utterance content of can be accurately collected.

(Embodiment 4)
The diagnostic device according to the fourth embodiment is characterized by diagnosing dementia using text data without performing voice recognition. FIG. 10 is a block diagram showing the overall configuration of diagnostic apparatus 1C according to Embodiment 4 of the present invention. In this embodiment, the same components as those in the first to third embodiments are designated by the same reference numerals and the description thereof will be omitted.

  The diagnostic device 1C is different from the diagnostic device 1B in that a scanner 910 is newly provided, and the USB memory 930, the IC recorder 940, and the voice recognition server 200 are omitted.

  The scanner 910 optically reads the document in which the utterance content of the diagnosis target person has been transcribed or the document written by the diagnosis target person, performs text recognition processing on the read document, generates text data, and causes the PC 920 to perform the text data generation. Output. Here, as the document in which the content of speech of the person to be diagnosed is transcribed, for example, the content of speech that the person to be diagnosed uttered through a dialogue with a doctor or the like or a message displayed on the display unit 921 is transcribed on a paper by the doctor or the like. Documents may be adopted.

  Further, as the document written by the diagnosis target person, for example, a document directly written on the paper by the diagnosis target person through a message displayed on the display unit 921 or a dialogue with a doctor or the like may be adopted.

  The USB interface 923 acquires the text data generated by the scanner 910.

  FIG. 11 is a flowchart showing an example of processing of the diagnostic device 1C according to the fourth embodiment of the present invention. First, in step S1101, the PC 920 acquires text data from the scanner 910. Next, the communication unit 922 transmits the acquired text data to the diagnosis server 300 via the network NT (S1102). The processing of S1111 to S1117 of FIG. 11 is the same as the processing of S301 to S307 of FIG. 3, and the processing of S1103 and S1104 of FIG. 11 is the same as the processing of S103 and S104 of FIG.

  As described above, in the diagnostic device 1C according to the fourth embodiment, since the text data is acquired using the scanner 910, the text data can be acquired without performing voice recognition, and the processing load of the entire system is reduced. it can.

X threshold 1,1A,1B,1C diagnostic device 100 portable terminal 101,410 sound collecting unit 102 touch panel 103 control unit 104,202,304,922 communication unit 200 speech recognition server 201,422 speech recognition unit 300 diagnostic server 301,423 Compression unit 302, 424 Compression ratio calculation unit 303 Determination unit 412 Display unit 420 Processing unit 421 Acquisition unit 425 Determination unit 910 Scanner 920 PC
921 Display unit 923 USB interface 930 USB memory 940 IC recorder

Claims (8)

A diagnostic device for diagnosing a disease in which language appears
An acquisition unit that acquires text data indicating the utterance content of the person or the document transcribed by the person,
A compression unit that reversibly compresses the text data acquired by the acquisition unit by a data compression method in which the compression rate increases as the number of repetitions of the same word increases ,
A compression rate calculation section for calculating the compression rate of the text data losslessly compressed by the compression section;
A diagnostic device comprising: a determination unit that determines whether or not the person has a suspicion of the disease based on the compression ratio calculated by the compression ratio calculation unit. A sound pickup unit that picks up the voice uttered by the person,
The diagnostic device according to claim 1, further comprising: a voice recognition unit configured to convert the collected voice into the text data and output the voice to the acquisition unit by voice-recognizing the collected voice.   3. The diagnostic apparatus according to claim 1, wherein the acquisition unit acquires the text data by recognizing a document in which the utterance content of the person is described or a document transcribed by the person as text.   The diagnostic device according to claim 1, wherein the determination unit determines that the person has the disease when the compression rate is higher than a predetermined threshold value.   The diagnostic device according to claim 1, wherein the determination unit determines that the degree of the disease is higher as the compression rate is higher.   The diagnostic apparatus according to claim 1, wherein the disease includes dementia, brain disease, and mental disorder. A method for controlling a diagnostic device for diagnosing a disease in which a symptom appears in a language,
By the diagnostic device,
Get the text data indicating the speech content or document which the person has transcript of human product,
The acquired text data is losslessly compressed by a data compression method in which the compression rate increases as the number of repetitions of the same word increases ,
Calculating the compression rate of the reversibly compressed text data,
A method for controlling a diagnostic device, which compares the calculated compression rate with a predetermined threshold value . A diagnostic program for diagnosing a disease that manifests in language,
A diagnostic program that causes a computer to function as each unit included in the diagnostic device according to claim 1.

Images