A cloud computing-based intelligent hearing and language rehabilitation platform

Abstract

The present invention discloses a cloud computing-based hearing and language intelligent rehabilitation platform, comprising a data acquisition module, a retrieval module, a content integration module and a rehabilitation test module. The data acquisition module is used to obtain relevant information of recording content; the retrieval module is used to retrieve the content information collected by the data acquisition module; the content integration module is used to integrate the retrieval content of the retrieval module; and the rehabilitation test module is used to test the rehabilitation status of patients. The present invention analyzes and integrates the recording content stored in a database, screens out the recording content suitable for rehabilitation testing of hearing-impaired patients, analyzes the rehabilitation status of patients through the rehabilitation test module, accurately obtains the hearing ability of patients, reduces the diagnostic burden of medical personnel, improves the daily level of test content, and minimizes the inaccuracy of manual diagnosis and treatment monitoring. The present invention has the characteristics of strong language optimization ability and high detection accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of hearing ability rehabilitation diagnosis, and in particular to a hearing and language intelligent rehabilitation platform based on cloud computing. Background Art

[0002] Statistics from the World Health Organization show that hearing loss has become the world's largest and growing sensory organ disease. For these hearing-impaired patients, cochlear implants are undoubtedly their ears for listening to the world. Hearing-impaired patients need extremely rich listening input to improve their speech, understanding and cognition. Brain development is crucial to the entire rehabilitation process. Therefore, with the rapid development of medical technology and doctors' clinical technical means, hearing loss is no longer just a problem of hearing aids.

[0003] The hearing impairment of many hearing-impaired people is manifested in their poor ability to understand the language they hear. In the prior art, there are medical systems that assist hearing-impaired people in rehabilitation testing. However, since the hearing test of the rehabilitation testing system is abstract and lacks reference, it is impossible to subdivide the test results of hearing-impaired patients through the results, resulting in a lack of accuracy in rehabilitation testing. At the same time, the pursuit of systematization in the rehabilitation testing system has resulted in overly rigid hearing test scenarios and test content, which are significantly different from daily life conversations, further reducing the accuracy of hearing rehabilitation testing for patients. Therefore, it is necessary to design a cloud-based hearing and language intelligent rehabilitation platform with strong language optimization capabilities and high detection accuracy. Summary of the invention

[0004] The purpose of the present invention is to provide a cloud computing-based hearing and language intelligent rehabilitation platform to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a cloud computing-based hearing and language intelligent rehabilitation platform, comprising a data acquisition module, a retrieval module, a content integration module and a rehabilitation test module, wherein the data acquisition module is used to obtain relevant information of the recorded content; the retrieval module is used to retrieve the content information collected by the data acquisition module; the content integration module is used to integrate the retrieval content of the retrieval module; and the rehabilitation test module is used to test the rehabilitation status of the patient.

[0006] According to the above technical solution, the content integration module includes a relationship calibration module and a speech recognition module, the relationship calibration module is used to calibrate the recorded content; the speech recognition module is used to convert the recorded sound into language content.

[0007] According to the above technical solution, the relationship calibration module further includes a first calibration submodule and a second calibration submodule. The first calibration submodule is used to calibrate based on the frequency of occurrence of names in the recording, and further calibrate the overlapping names in the recording whose frequency of occurrence reaches the required level; the second calibration submodule is used to screen and calibrate the "relationship set group" obtained by the system based on the overlapping names in the recording.

[0008] According to the above technical solution, the rehabilitation test module includes a test module and a test result analysis module. The test module is used to test the patient's hearing rehabilitation status; the test result analysis module is used to analyze the patient's hearing ability through the test results.

[0009] According to the above technical solution, the operation method of the hearing and language intelligent rehabilitation platform mainly includes the following steps:

[0010] Step S1: The data collection module obtains the information that can be authorized to extract recordings from databases in different fields through various channels, and then divides each recording information into a recording content every three minutes;

[0011] Step S2: the first calibration submodule performs calibration based on the frequency of occurrence of the names in the recording, and further calibrates the names that overlap in the recording and whose frequency of occurrence reaches the required level;

[0012] Step S3: The second calibration submodule is used to screen and calibrate the "relationship set group" obtained by the system according to the overlapping names in the recording;

[0013] Step S4: The rehabilitation test module performs a rehabilitation test on the patient;

[0014] Step S5: The system obtains the patient's hearing ability through the rehabilitation test results and outputs the results to medical personnel.

[0015] According to the above technical solution, step S2 further includes:

[0016] Step S21: The speech recognition module obtains the specific language content of the recording by recognizing the recording, and the key point checking submodule screens the names of the subject and the object in the language content, predicts and expands the pronouns therein, obtains the occurrence frequency of the target name and its pronouns in all the names, and extracts the names with a frequency of more than 20% in the recording names;

[0017] Step S22: If the number of names extracted from the recording is greater than 5, the current recording is stored in the rehabilitation test module; if the number of names extracted from the recording is greater than 5 at least two, the currently appearing name is stored as a "name to be synchronized"; otherwise the recording content is deleted.

[0018] According to the above technical solution, step S3 further includes:

[0019] Step S31: the speech recognition module sets the word segments with the same target name and pronoun in the recording content as a label segment in the form of "target name + part", and sets the label segment to a color;

[0020] Step S32: the speech recognition module sets the next different target name and pronoun appearing in the recorded language content as a label segment in the form of "target name + part", and sets the label segment to another color;

[0021] Step S33: When the target name that appears is a "name to be synchronized", a new "relationship set group" is created and the "name to be synchronized" is copied into the group, and all subsequent "names to be synchronized" that appear are copied into the group according to step S2. When the recording is finished, all "names to be synchronized" in the "relationship set group" are connected in series and it is determined through the data in the database whether the "relationship set group" can form a complete relationship logic chain. The incomplete relationship logic chain is supplemented according to the existing data. After the relationship logic chain is complete, all "names to be synchronized" in the current "relationship set group" are updated to "synchronized names".

[0022] Step S34: When the target name that appears is a "name to be synchronized" and there is a new "name to be synchronized" associated with it in step S2, a new "relationship set group" is created and all the "names to be synchronized" in step S2 are copied into the new group, and all subsequent "names to be synchronized" that appear in step S2 are copied into the new group. When the recording is finished, all the "names to be synchronized" in the "relationship set group" are connected in series and it is determined through the data in the database whether the new "relationship set group" can form a complete relationship logic chain, and the incomplete relationship logic chain is supplemented according to the existing data. After the relationship logic chain is complete, all the "names to be synchronized" in the new "relationship set group" are updated to "synchronized names";

[0023] Step S35: Repeat the above steps to divide the recorded language content into n different label segments. The third calibration submodule compares the different "overlapping name groups" with the overlapping color parts of the label segments with all the "relationship set groups", and matches the two groups one by one according to the criterion of the most overlapping label names in the two groups, and deletes the "relationship set groups" with more than 50% non-overlapping names in the corresponding groups and the "relationship set groups" that have not been matched.

[0024] According to the above technical solution, step S4 further includes:

[0025] Step S41: The content integration module reorders the recording content according to the narration order in the recording, and integrates the connection relationship of all names in the recording through the remaining "relationship set groups", and finally generates a rehabilitation test recording for the hearing-impaired patient and tests the patient through the rehabilitation test module:

[0026] Step S42: The system uses the content integration module to classify the recorded content into six difficulty levels according to the standards:

[0027] The first recording is marked as "difficulty level 1 test", and the generated test questions are (1) + (2) + (3) + (4) + (5), where (1), (2), (3), (4), and (5) are the names of five different subjects obtained by the system in the recording, and three distractor options are provided when the correct answer is provided. The second recording is marked as "difficulty level 2 test", and the generated test questions are (1) + (11) + (12) + (13), where (1) is the name of the subject obtained by the system in the recording, and two distractor options are provided when the correct answer is provided; (11), (12), and (13) are The system obtains three event summaries from the recording, and provides six distractor options while providing the correct answer. The third recording content is marked as "difficulty level 3 test", and the generated test questions are (11) + (12) + (13) + (111) + (122) + (132), where (11), (12), and (13) are the names of three different subjects obtained by the system from the recording, and provide two distractor options while providing the correct answer. (111), (122), and (132) are the event summaries of the subjects obtained by the system from the recording, and provide two distractor options while providing the correct answer.

[0028] The fourth recording is marked as "difficulty level 4 test", and the generated test questions are (11) + (12) + (13) + (14) + (1112) + (1213) + (1314) + (1411), where (11), (12), and (13) are four different subject names obtained by the system in the recording, and two interference options are provided for each of the four subjects when the correct answer is provided; (1112), (1213), (1314), and (1411) are four event summaries of two subjects obtained by the system in the recording, and two interference options are provided for each of the four subjects when the correct answer is provided; Step S43: The patient starts the test from difficulty level 1, performs two tests at the current difficulty, and performs the following effects:

[0029] If the sum of the correct rates of the two test results is higher than 160%, the next test of difficulty will be started;

[0030] If the sum of the accuracy of the two test results is less than 100%, restart the test of the previous difficulty;

[0031] Otherwise, the hearing ability of the current patient's hearing rehabilitation index corresponding to the difficulty level is determined;

[0032] When the patient is taking the "difficulty level 1 test", if the sum of the correct rates of the two test results is less than 100%, the patient's hearing rehabilitation index is judged to correspond to difficulty level 0;

[0033] When the patient is taking the "difficulty level 6 test", if the sum of the correct rates of the two test results is higher than 160%, it is judged that the patient's current hearing rehabilitation index corresponds to difficulty level 5;

[0034] If the system reaches a certain difficulty level for the third time after the above steps, the hearing ability of the current patient's hearing rehabilitation index corresponding to the difficulty level is directly determined. According to the above technical solution, step S5 further includes:

[0035] Step S51: After the test is completed, the rehabilitation test module obtains the hearing ability corresponding to each difficulty level:

[0036] Difficulty level 0: The patient has a weak ability to recognize sounds and severe memory impairment;

[0037] Difficulty level 1: The patient has basic ability to recognize sounds, but is weak in extracting key information from language;

[0038] Difficulty level 2: The patient has the ability to extract key information from the language, but cannot identify intertwining relationships;

[0039] Difficulty level 3: The patient can identify simple correspondences between various information in the language, but cannot identify

[0040] Difficulty level 4: The patient can identify the more complex correspondence between the various information in the language;

[0041] Difficulty level 5: The patient can sort out the complex correspondence between various information and obtain accurate results;

[0042] Step S51: Medical personnel treat the patient based on the patient's hearing condition and their own diagnosis results obtained by the system.

[0043] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: the present invention, through the retrieval module and the content integration module, analyzes and integrates the recording content stored in the database, screens out the recording content suitable for rehabilitation testing of hearing-impaired patients, and analyzes the patient's rehabilitation status through the rehabilitation test module, accurately obtains the patient's hearing ability, and conducts diagnosis and treatment in a timely manner, which reduces the diagnostic burden of medical personnel while improving the daily level of test content, and minimizes the inaccuracy of manual diagnosis and treatment monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0045] Figure 1 It is a schematic diagram of the system module composition of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] See also Figure 1 The present invention provides a technical solution: a cloud computing-based hearing and language intelligent rehabilitation platform, comprising:

[0048] Data collection module, retrieval module, content integration module and rehabilitation test module, the data collection module is used to obtain relevant information of the recording content; the retrieval module is used to retrieve the content information collected by the data collection module; the content integration module is used to integrate the retrieval content of the retrieval module; the rehabilitation test module is used to test the rehabilitation status of the patient.

[0049] The present invention analyzes and integrates the recorded contents stored in the database through the retrieval module and the content integration module, screens out the recorded contents suitable for rehabilitation testing of hearing-impaired patients, and analyzes the rehabilitation status of the patients through the rehabilitation test module, accurately obtains the hearing ability of the patients, and performs diagnosis and treatment in a timely manner, thereby reducing the diagnostic burden of medical personnel, improving the daily level of test content, and minimizing the inaccuracy of manual diagnosis and treatment monitoring.

[0050] The content integration module includes a relationship calibration module and a speech recognition module. The relationship calibration module is used to calibrate the recorded content; the speech recognition module is used to convert the recorded sound into language content.

[0051] The relationship calibration module further includes a first calibration submodule and a second calibration submodule. The first calibration submodule is used to calibrate based on the frequency of occurrence of names in the recording, and further calibrate the overlapping names in the recording whose frequency of occurrence reaches the required level; the second calibration submodule is used to screen and calibrate the "relationship set group" obtained by the system based on the overlapping names in the recording.

[0052] The rehabilitation test module includes a test module and a test result analysis module. The test module is used to test the patient's hearing rehabilitation status; the test result analysis module is used to analyze the patient's hearing ability through the test results.

[0053] In a preferred embodiment, the operation method of the hearing and language intelligent rehabilitation platform mainly includes the following steps:

[0054] Step S1: The data collection module obtains the information that can be authorized to extract recordings from databases in different fields through various channels, and then divides each recording information into a recording content every three minutes;

[0055] Step S2: the first calibration submodule performs calibration based on the frequency of occurrence of the names in the recording, and further calibrates the names that overlap in the recording and whose frequency of occurrence reaches the required level;

[0056] Step S3: The second calibration submodule is used to screen and calibrate the "relationship set group" obtained by the system according to the overlapping names in the recording;

[0057] Step S4: The rehabilitation test module performs a rehabilitation test on the patient;

[0058] Step S5: The system obtains the patient's hearing ability through the rehabilitation test results and outputs the results to medical personnel.

[0059] In this embodiment, step S2 further includes:

[0060] Step S21: The speech recognition module obtains the specific language content of the recording by recognizing the recording, and the key point checking submodule screens the names of the subject and object in the language content, predicts and expands the pronouns therein, obtains the frequency of occurrence of the target name and its pronouns in all names, and extracts the names with a frequency of occurrence exceeding 20% ​​in the recording names; when there are many names in the recording content, the main subject of the recording content can be screened out by extracting the names with a frequency of occurrence exceeding 20% ​​in the recording names, which is conducive to increasing the content concentration of the rehabilitation content and improving the accuracy of the test.

[0061] Step S22: If the number of names extracted from the recording is greater than 5, the current recording is stored in the rehabilitation test module; if the number of names extracted from the recording is greater than 5, at least two of them are greater than 5, the currently appearing name is stored as a "name to be synchronized"; otherwise, the recording content is deleted. When the subject of the recording content is small and the event content is large, the content line is single and the main content of the event is simple; when the subject of the recording content is small and the event content is large, there are multiple content lines and the main content of the event is complex.

[0062] In this embodiment, step S3 further includes:

[0063] Step S31: the speech recognition module sets the word segments with the same target name and pronoun in the recording content as a label segment in the form of "target name + part", and sets the label segment to a color;

[0064] Step S32: the speech recognition module sets the next different target name and pronoun appearing in the recorded language content as a label segment in the form of "target name + part", and sets the label segment to another color;

[0065] Step S33: When the target name that appears is a "name to be synchronized", a new "relationship set group" is created and the "name to be synchronized" is copied into the group, and all subsequent "names to be synchronized" that appear are copied into the group according to step S2. When the recording is finished, all "names to be synchronized" in the "relationship set group" are connected in series and it is determined through the data in the database whether the "relationship set group" can form a complete relationship logic chain. The incomplete relationship logic chain is supplemented according to the existing data. After the relationship logic chain is complete, all "names to be synchronized" in the current "relationship set group" are updated to "synchronized names".

[0066] Step S34: When the target name that appears is a "name to be synchronized" and there is a new "name to be synchronized" associated with it in step S2, a new "relationship set group" is created and all the "names to be synchronized" in step S2 are copied into the new group, and all subsequent "names to be synchronized" that appear in step S2 are copied into the new group. When the recording is finished, all the "names to be synchronized" in the "relationship set group" are connected in series and it is determined through the data in the database whether the new "relationship set group" can form a complete relationship logic chain, and the incomplete relationship logic chain is supplemented according to the existing data. After the relationship logic chain is complete, all the "names to be synchronized" in the new "relationship set group" are updated to "synchronized names";

[0067] Step S35: Repeat the above steps to divide the recorded language content into n different label segments. The third calibration submodule compares the different "overlapping name groups" with the overlapping color parts of the label segments with all the "relationship set groups", and matches the two groups one by one according to the criterion of the most overlapping label names in the two groups, and deletes the "relationship set groups" with more than 50% non-overlapping names in the corresponding groups and the "relationship set groups" that have not been matched.

[0068] Through the first calibration submodule and the second calibration submodule, the recorded data in the database are identified and screened according to the test requirements, providing a favorable basis for subsequent rehabilitation tests on hearing-impaired patients. At the same time, due to the randomness of the extracted data, the content scenarios of the rehabilitation test are highly random, thereby improving the accuracy of the rehabilitation test.

[0069] In this embodiment, step S4 further includes:

[0070] Step S41: The content integration module reorders the recording content according to the narration order in the recording, and integrates the connection relationship of all names in the recording through the remaining "relationship set groups", and finally generates a rehabilitation test recording for the hearing-impaired patient and tests the patient through the rehabilitation test module:

[0071] Step S42: The system uses the content integration module to classify the recorded content into six difficulty levels according to the standards:

[0072] The first recording is marked as "difficulty level 1 test", and the generated test questions are (1) + (2) + (3) + (4) + (5), where (1), (2), (3), (4), and (5) are the names of five different subjects obtained by the system in the recording. Three distractor options are provided along with the correct answer. All the options are subjects. This level of test only requires information recognition, and does not require information screening and extraction.

[0073] The second recording content is marked as "difficulty level 2 test", and the generated test questions are (1) + (11) + (12) + (13), where (1) is the subject name obtained by the system in the recording, and two distractor options are provided when the correct answer is provided; (11), (12), and (13) are three event summaries obtained by the system in the recording, and six distractor options are provided when the correct answer is provided; the options are single subject + event. This level focuses on the event of the target subject while identifying information, and examines the patient's basic ability to extract language information;

[0074] The third recording content is marked as "difficulty level 3 test", and the generated test questions are (11) + (12) + (13) + (111) + (122) + (132), where (11), (12), and (13) are three different subject names obtained by the system in the recording, and two interference options are provided for each of them when the correct answer is provided; (111), (122), and (132) are event summaries of the subjects obtained by the system in the recording, and two interference options are provided for each of them when the correct answer is provided; the options are multiple subjects + multiple times. This level adds multiple target subjects to test the patient's ability to identify each subject and the corresponding event information;

[0075] The content of the fourth recording is marked as "Difficulty Level 4 Test", and the generated test questions are (11) + (12) + (13) + (14) + (1112) + (1213) + (1314) + (1411), where (11), (12), and (13) are four different subject names obtained by the system in the recording, and two interference options are provided for each of them when the correct answer is provided; (1112), (1213), (1314), and (1411) are four event summaries obtained by the system in the recording, in which each of the two subjects is involved, and two interference options are provided for each of them when the correct answer is provided; the options are multiple subjects + intertwined relationships of events. This level requires the patient to have the ability to identify and integrate the relationship information between subjects, between events and subjects, and between events.

[0076] Step S43: The patient starts the test from difficulty level 1, performs two tests at the current difficulty level, and performs the following effects:

[0077] If the sum of the correct rates of the two test results is higher than 160%, the next test of difficulty will be started;

[0078] If the sum of the accuracy of the two test results is less than 100%, restart the test of the previous difficulty;

[0079] Otherwise, the hearing ability of the current patient's hearing rehabilitation index corresponding to the difficulty level is determined;

[0080] When the patient is taking the "difficulty level 1 test", if the sum of the correct rates of the two test results is less than 100%, the patient's hearing rehabilitation index is judged to correspond to difficulty level 0;

[0081] When the patient is taking the "difficulty level 6 test", if the sum of the correct rates of the two test results is higher than 160%, it is judged that the patient's current hearing rehabilitation index corresponds to difficulty level 5;

[0082] If the system reaches a certain difficulty level for the third time after the above steps, the hearing ability of the current patient's hearing rehabilitation index corresponding to the difficulty level is directly judged. When the test reaches a certain difficulty level for the third time, since the hearing rehabilitation test results of the patient at each level show a large contrast, the patient's actual hearing ability is evaluated through the repetition of the test level, which reduces the load of the system operation and improves the efficiency of the rehabilitation test.

[0083] In this embodiment, step S5 further includes:

[0084] Step S51: After the test is completed, the rehabilitation test module obtains the hearing ability corresponding to each difficulty level:

[0085] Difficulty level 0: The patient has a weak ability to recognize sounds and severe memory impairment;

[0086] Difficulty level 1: The patient has basic ability to recognize sounds, but is weak in extracting key information from language;

[0087] Difficulty level 2: The patient has the ability to extract key information from the language, but cannot identify intertwining relationships;

[0088] Difficulty level 3: The patient can identify simple correspondences between various information in the language, but cannot recognize them;

[0089] Difficulty level 4: The patient can identify the more complex correspondence between the various information in the language;

[0090] Difficulty level 5: The patient can sort out the complex correspondence between various information and obtain accurate results;

[0091] The rehabilitation test module can differentiate and analyze the hearing abilities of patients at each test level, solving the problem of being unable to subdivide the test results of hearing-impaired patients. This improves the accuracy of the patient's hearing rehabilitation test and reduces the diagnostic burden on medical staff.

[0092] Step S51: Medical personnel treat the patient based on the patient's hearing condition and their own diagnosis results obtained by the system.

[0093] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0094] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A cloud computing-based hearing and language intelligent rehabilitation platform, characterized by: It includes a data acquisition module, a retrieval module, a content integration module and a rehabilitation test module. The data acquisition module is used to obtain relevant information of the recording content; the retrieval module is used to retrieve the content information collected by the data acquisition module; the content integration module is used to integrate the retrieved content of the retrieval module; and the rehabilitation test module is used to test the rehabilitation status of the patient; The content integration module includes a relationship calibration module and a speech recognition module, wherein the relationship calibration module is used to calibrate the recorded content; and the speech recognition module is used to convert the recorded sound into language content. The relationship calibration module further includes a first calibration submodule and a second calibration submodule, wherein the first calibration submodule is used to calibrate by the frequency of occurrence of names in the recording, and further calibrate the names that overlap in the recording and whose frequency of occurrence reaches the required level; the second calibration submodule is used to screen and calibrate the "relationship set group" obtained by the system by the names that overlap in the recording; The rehabilitation test module includes a test module and a test result analysis module, and the test module is used to test the hearing rehabilitation status of the patient; The test result analysis module is used to analyze the patient's hearing ability through the test results; The operation method of the hearing and language intelligent rehabilitation platform mainly includes the following steps: Step S1: The data collection module obtains the information that can be authorized to extract recordings from databases in different fields through various channels, and then divides each recording information into a recording content every three minutes; Step S2: the first calibration submodule performs calibration based on the frequency of occurrence of the names in the recording, and further calibrates the names that overlap in the recording and whose frequency of occurrence reaches the required level; Step S3: The second calibration submodule is used to screen and calibrate the "relationship set group" obtained by the system according to the overlapping names in the recording; Step S4: The rehabilitation test module performs a rehabilitation test on the patient; Step S5: The system obtains the patient's hearing ability through the rehabilitation test results and outputs the results to medical personnel; The step S2 further comprises: Step S21: The speech recognition module obtains the specific language content of the recording by recognizing the recording, and the key point checking submodule screens the names of the subject and the object in the language content, predicts and expands the pronouns therein, obtains the occurrence frequency of the target name and its pronouns in all the names, and extracts the names with a frequency of more than 20% in the recording names; Step S22: if the number of names extracted from the recording is greater than 5, the current recording is stored in the rehabilitation test module; if the number of names extracted from the recording is greater than 5 at least two, the currently appearing name is stored as "name to be synchronized"; otherwise, the recording content is deleted; The step S3 further comprises: Step S31: the speech recognition module sets the word segments with the same target name and pronoun in the recording content as a label segment in the form of "target name + part", and sets the label segment to a color; Step S32: the speech recognition module sets the next different target name and pronoun appearing in the recorded language content as a label segment in the form of "target name + part", and sets the label segment to another color; Step S33: When the target name that appears is a "name to be synchronized", a new "relationship set group" is created and the "name to be synchronized" is copied into the group, and all subsequent "names to be synchronized" that appear are copied into the group according to step S2. When the recording is finished, all "names to be synchronized" in the "relationship set group" are connected in series and it is determined through the data in the database whether the "relationship set group" can form a complete relationship logic chain. The incomplete relationship logic chain is supplemented according to the existing data. After the relationship logic chain is complete, all "names to be synchronized" in the current "relationship set group" are updated to "synchronized names"; Step S34: When the target name that appears is a "name to be synchronized" and there is a new "name to be synchronized" associated with it in step S2, a new "relationship set group" is created and all the "names to be synchronized" in step S2 are copied into the new group, and all subsequent "names to be synchronized" that appear in step S2 are copied into the new group. When the recording is finished, all the "names to be synchronized" in the "relationship set group" are connected in series and it is determined through the data in the database whether the new "relationship set group" can form a complete relationship logic chain. The incomplete relationship logic chain is supplemented according to the existing data. After the relationship logic chain is complete, all the "names to be synchronized" in the new "relationship set group" are updated to "synchronized names"; Step S35: Repeat the above steps to divide the recorded language content into n different label segments. The third calibration submodule compares the different "overlapping name groups" of the color-overlapping parts of the label segments with all the "relationship set groups", and matches the two groups one by one according to the criterion of the most overlapping label names in the two groups, and deletes the "relationship set groups" with non-overlapping names exceeding 50% in the corresponding groups and the "relationship set groups" that have not been matched; The step S4 further comprises: Step S41: The content integration module reorders the recording content according to the narration order in the recording, and integrates the connection relationship of all names in the recording through the remaining "relationship set groups", and finally generates a rehabilitation test recording for the hearing-impaired patient and tests the patient through the rehabilitation test module: Step S42: The system uses the content integration module to classify the recorded content into six difficulty levels according to the standards: The first recording is marked as "difficulty level 1 test" and the generated test questions are (1) + (2) + (3) + (4) + (5), where (1), (2), (3), (4), and (5) are the names of five different subjects obtained by the system in the recording. The correct answer is provided along with three distractor options. The second recording content is marked as "difficulty level 2 test", and the generated test questions are (1) + (11) + (12) + (13), where (1) is the subject name obtained by the system in the recording, and provides two distractor options while providing the correct answer; (11), (12), and (13) are three event summaries obtained by the system in the recording, and provide six distractor options while providing the correct answer; The third recording content is marked as "difficulty level 3 test", and the generated test questions are (11) + (12) + (13) + (111) + (122) + (132), where (11), (12), and (13) are the names of three different subjects obtained by the system in the recording, and two distractor options are provided for each of them when the correct answer is provided; (111), (122), and (132) are the event summaries of the subjects obtained by the system in the recording, and two distractor options are provided for each of them when the correct answer is provided; The fourth recording is marked as "difficulty level 4 test", and the generated test questions are (11) + (12) + (13) + (14) + (1112) + (1213) + (1314) + (1411), where (11), (12), and (13) are four different subject names obtained by the system in the recording, and two interference options are provided for each of them when a correct answer is provided; (1112), (1213), (1314), and (1411) are four event summaries of two subjects obtained by the system in the recording, and two interference options are provided for each of them when a correct answer is provided; Step S43: The patient starts the test from difficulty level 1, performs two tests at the current difficulty, and performs the following effects: If the sum of the correct rates of the two test results is higher than 160%, the next test of difficulty will be started; If the sum of the accuracy of the two test results is less than 100%, restart the test of the previous difficulty; Otherwise, the hearing ability of the current patient's hearing rehabilitation index corresponding to the difficulty level is determined; When the patient is taking the "difficulty level 1 test", if the sum of the correct rates of the two test results is less than 100%, the patient's hearing rehabilitation index is judged to correspond to difficulty level 0; When the patient is taking the "difficulty level 4 test", if the sum of the correct rates of the two test results is higher than 160%, it is judged that the current patient's hearing rehabilitation index corresponds to difficulty level 5; If the system reaches a certain difficulty level for the third time after the above steps, the hearing ability of the current patient's hearing rehabilitation index corresponding to the difficulty level is directly determined.

2. The cloud computing-based hearing and language intelligent rehabilitation platform according to claim 1, characterized in that: The step S5 further comprises: Step S51: After the test is completed, the rehabilitation test module obtains the hearing ability corresponding to each difficulty level: Difficulty level 0: The patient has a weak ability to recognize sounds and severe memory impairment; Difficulty level 1: The patient has basic ability to recognize sounds, but is weak in extracting key information from language; Difficulty level 2: The patient has the ability to extract key information from the language, but cannot identify intertwining relationships; Difficulty level 3: The patient can identify simple correspondences between various information in the language, but cannot identify Difficulty level 4: The patient can identify the more complex correspondence between the various information in the language; Difficulty level 5: The patient can sort out the complex correspondence between various information and obtain accurate results; Step S51: Medical personnel treat the patient based on the patient's hearing condition and their own diagnosis results obtained by the system.

Images