English phoneme personalized correction system and method based on acoustic feature analysis

Abstract

The invention relates to the technical field of English phoneme personalized correction systems, in particular to an English phoneme personalized correction system and method based on acoustic feature analysis. Comprising a learner identity recognition module, a learner pronunciation information input module, a learner pronunciation information analysis module, a learner pronunciation difficulty point evaluation module, a learner personalized correction strategy generation module, a learner pronunciation difficulty correction module, a voice model evaluation module and a learning effect feedback module which are connected in sequence. According to the technical scheme of the invention, voice learners are served from multiple angles and in all directions from links such as information acquisition, problem diagnosis, strategy making, process monitoring and effect evaluation, and a set of intelligent and personalized voice learning solution is provided for the learners through innovative fusion of acoustic voice analysis and artificial intelligence technology. The method is of great significance in improving the voice level of English learners and promoting cross-cultural communication.

Description

Technical Field

[0001] The present invention relates to the technical field of English phoneme personalized correction system, in particular to an English phoneme personalized correction system and method based on acoustic feature analysis. Background Art

[0002] With the deepening of economic globalization, the importance of English as a common language has become increasingly prominent. However, due to the influence of factors such as negative transfer from their mother tongue, non-native English learners often encounter many difficulties in the process of learning English, especially in pronunciation. Inaccurate pronunciation not only affects the expression of phonetic semantics, but also brings obstacles to cross-cultural communication. Therefore, how to help English learners overcome negative transfer from their mother tongue and improve their English pronunciation has become an urgent problem to be solved in the field of English teaching.

[0003] At present, common English phonetics teaching methods mainly include listening to recordings, imitating, and face-to-face guidance from phonetics teachers. Although these methods can help learners improve their pronunciation to a certain extent, they still have some shortcomings. For example, listening to recordings, imitating, and imitating lack interactivity and pertinence, and learners find it difficult to obtain timely feedback and personalized guidance; face-to-face guidance from phonetics teachers is limited by objective conditions such as time and place, and cannot meet learners' needs to learn anytime and anywhere; moreover, neither recordings nor teachers can accurately assess learners' pronunciation problems, and it is difficult to provide targeted correction plans.

[0004] In addition, although some existing computer-assisted speech learning systems can realize human-computer interaction and automatic speech recognition, their functions are relatively simple and it is difficult to comprehensively evaluate learners' speech ability. Most of these systems use a general speech recognition model and cannot fully consider the impact of negative transfer from the mother tongue; they only focus on the accuracy of pronunciation and ignore suprasegmental features such as rhythm and intonation of speech; and they cannot generate personalized correction plans based on individual differences of learners. In general, the existing English speech teaching methods and tools are not intelligent and personalized enough, and it is difficult to effectively help learners overcome pronunciation barriers and improve their speech level. Summary of the invention

[0005] In view of the above technical problems, the present invention proposes an English phoneme personalized correction system and method based on acoustic feature analysis. The system makes full use of acoustic speech analysis technology, comprehensively collects and analyzes learners' pronunciation data, considers influencing factors such as negative transfer of the mother tongue, and evaluates learners' speech mastery and weak links from multiple angles; on this basis, from multiple dimensions such as pronunciation training content, training methods, and tutors, a personalized correction plan is tailored for learners, and dynamically adjusted and optimized according to the learning effect. Through the application of this system, the speech correction efficiency and learning effect of English learners can be maximized.

[0006] The present invention proposes an English phoneme personalized correction system based on acoustic feature analysis, comprising a learner identity recognition module, a learner pronunciation information input module, a learner pronunciation information analysis module, a learner pronunciation difficulty point evaluation module, a learner personalized correction strategy generation module, a learner difficult pronunciation correction module, a speech model evaluation module and a learning effect feedback module, wherein the learner identity recognition module is used to identify the learner's native language type and collect learner information including the learner's age, gender and English language background, the learner pronunciation information input module is used to receive pronunciation information submitted by the learner and have the learner mark the target phoneme to be learned, the learner pronunciation information analysis module is used to calculate the difference data of the acoustic features of the learner's pronunciation and the acoustic features of the target phoneme after normalization processing and analyze the learner's mastery of different target phonemes based on the data, and the learner pronunciation information analysis module is used to calculate the difference data of the acoustic features of the learner's pronunciation and the acoustic features of the target phoneme after normalization processing and analyze the learner's mastery of different target phonemes based on the data, and the learner pronunciation information analysis module is used to calculate the difference data of the acoustic features of the learner's pronunciation and the target phoneme after normalization processing and analyze the learner's mastery of different target phonemes based on the data, and the learner pronunciation information analysis module is used to calculate the difference data of the acoustic features of the learner's pronunciation and the target phoneme The learner pronunciation difficulty assessment module is used to extract and analyze the learner's pronunciation features, timing features and rule features when learning a second language, and to assess the learner's pronunciation difficulties in combination with the learner's English language background, English level and the differential data output by the pronunciation information analysis module. The learner personalized correction strategy generation module is used to generate a personalized correction training plan including the type of corrected pronunciation, the number of exercises, the frequency of exercises, the length of exercises and the required learning instructors according to the pronunciation difficulties. The learner difficult pronunciation correction module is used to guide the learner to learn and correct difficult phonemes according to the personalized correction training plan. The speech model assessment module is used to quantitatively evaluate the learner's pronunciation according to a preset speech model. The learning effect feedback module is used to optimize the personalized correction training plan according to the quantitative evaluation results and adjust the corresponding learning instructors.

[0007] Preferably, the pronunciation information received by the learner pronunciation information input module includes video information, audio information and text information, wherein the video information and audio information record the learner's pronunciation practice process, and the text information contains voice text content. The learner pronunciation information input module is also used to receive the learner's annotations on the pronunciation information, and the annotations include the types of pronunciation practiced and the time intervals between different pronunciations.

[0008] Preferably, the learner pronunciation information analysis module includes an acoustic feature extraction unit, a feature comparison unit and a mastery calculation unit, wherein the acoustic feature extraction unit is used to extract acoustic features from the learner pronunciation and the target phonemes respectively, the feature comparison unit is used to normalize the extracted acoustic features and calculate difference data, and the mastery calculation unit is used to calculate the learner's mastery of different target phonemes based on the difference data.

[0009] Preferably, the learner pronunciation difficulty assessment module includes a phoneme contrast analysis unit, a rule analysis unit and a timing analysis unit, wherein the phoneme contrast analysis unit is used to contrast and analyze similar phonemes in the user's native language and English, the rule analysis unit is used to analyze similarity rules based on preset linguistic data, and the timing analysis unit is used to analyze the timing relationship between different phonemes.

[0010] Preferably, the preset linguistic materials include native language linguistic materials and second language learning materials, wherein the native language linguistic materials include the pronunciation and pronunciation rules of the native language, and the second language learning materials include the pronunciation and pronunciation rules of the second language.

[0011] The method for personalized correction of English phonemes based on acoustic feature analysis is applied to the system and comprises the following steps: collecting and recording the learner's native language type, age, gender, English language background and English proficiency information through the learner identity recognition module; receiving the learner's pronunciation information including video information, audio information and text information through the learner pronunciation information input module; performing acoustic feature analysis on the pronunciation information and calculating the difference data with the target phoneme through the learner pronunciation information analysis module; evaluating the learner's pronunciation difficulty points according to the difference data through the learner pronunciation difficulty point evaluation module; generating a personalized correction training plan through the learner personalized correction strategy generation module; executing the personalized correction training plan through the learner difficult pronunciation correction module; evaluating the learning effect through the speech model evaluation module; and optimizing the personalized correction training plan through the learning effect feedback module.

[0012] Preferably, the acoustic feature analysis includes pronunciation duration analysis, phoneme frequency analysis and pronunciation timing analysis, wherein the pronunciation duration analysis is used to analyze the difference in phoneme duration between the practice pronunciation and the target pronunciation, the phoneme frequency analysis is used to analyze the difference in phoneme frequency between the practice pronunciation and the target pronunciation, and the pronunciation timing analysis is used to analyze the difference in timing state between the practice pronunciation and the target pronunciation, and the timing state includes the previous phoneme, the next phoneme and the timing interval between phonemes.

[0013] Preferably, the execution of the personalized correction training program includes: matching an instructor according to the learner's pronunciation difficulties; performing pronunciation correction training according to the personalized correction training program; collecting pronunciation data during the training process; and evaluating the collected pronunciation data based on the speech model.

[0014] Preferably, the matching of the instructors includes: ranking the instructors according to pronunciation similarity; screening the instructors according to the evaluation of the teachers' teaching effectiveness; performing adaptability analysis according to the teachers' current workload; and selecting the most suitable instructor according to the above analysis results.

[0015] Preferably, the optimization of the personalized correction training program includes: analyzing the learning effect according to the speech model evaluation results; adjusting the training content and training parameters according to the learning effect; adjusting the instructor according to the learner feedback; and updating the pronunciation difficulty points according to the latest pronunciation data.

[0016] Compared with the prior art, the beneficial effects of the present invention are embodied in the following aspects:

[0017] First, in terms of learner pronunciation information collection, the present invention receives multimodal data including video, audio, and text through a learner pronunciation information input module, and the learner marks the pronunciation type and timing, which can comprehensively record the learner's pronunciation characteristics and overcome the deficiency of the single data dimension of the prior art.

[0018] Secondly, in terms of pronunciation problem diagnosis, the present invention comprehensively considers acoustic features such as voice, rhythm, and tone pattern, and evaluates the learner's mastery of different speech components from multiple angles through steps such as acoustic feature extraction, feature comparison, and mastery assessment. It also deeply analyzes the learner's pronunciation difficulties from the aspects of pronunciation features, timing features, and rule features, thus overcoming the superficial diagnosis limitations of existing technologies.

[0019] Thirdly, in terms of generating personalized correction strategies, the present invention makes full use of prior knowledge such as learners' language background, mastery level, pronunciation weaknesses, etc., and formulates targeted personalized correction plans from the dimensions of correction content, training methods, and teaching staff. Compared with the traditional "one-size-fits-all" approach, the present invention can teach students in accordance with their aptitude and improve the effectiveness of correction.

[0020] In addition, in terms of correction process monitoring, the present invention implements personalized correction plans through the learner's difficult pronunciation correction module, collects and analyzes training data in real time, uses the speech recognition model to objectively evaluate the training effect, combines the learner's subjective feedback, dynamically optimizes and adjusts the training parameters, forms a closed-loop feedback, and can maximize learning efficiency.

[0021] Finally, from the perspective of system architecture, the present invention adopts a modular design approach, with each module having a clear division of labor and close collaboration, which not only highlights the targeted nature of each link, but also takes into account the systematic nature of the overall process. It solves the connection problems of data collection, problem diagnosis, strategy formulation, and effect evaluation, and truly realizes personalized speech learning and correction in a closed loop throughout the entire process.

[0022] In summary, the personalized English phoneme correction system and method based on acoustic feature analysis provided by the present invention serve speech learners from multiple angles and in an all-round manner from the aspects of information collection, problem diagnosis, strategy formulation, process monitoring, and effect evaluation. Through the innovative integration of acoustic speech analysis and artificial intelligence technology, it provides learners with a set of intelligent and personalized speech learning solutions, which is of great significance to improving the speech level of English learners and promoting cross-cultural communication, and is a beneficial exploration in the field of language education. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 It is the overall logic block diagram of the system of the present invention.

[0024] Figure 2 This is the detailed process of the pronunciation information analysis module of the present invention.

[0025] Figure 3 It is a logic block diagram of the phoneme comparison and analysis unit of the present invention.

[0026] Figure 4 It is a logic block diagram of the rule analysis unit of the present invention.

[0027] Figure 5 It is a logic block diagram of the timing analysis unit of the present invention.

[0028] Figure 6 This is the correction training workflow of the present invention.

[0029] Figure 7 Feedback optimization process for the present invention. DETAILED DESCRIPTION

[0030] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects thereof are described in detail below in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0031] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0032] Example 1

[0033] See also Figure 1-7 The present invention relates to an English phoneme personalized correction system and method based on acoustic feature analysis.

[0034] The English phoneme personalized correction system based on acoustic feature analysis of the present invention comprises a learner identity recognition module 1, a learner pronunciation information input module 2, a learner pronunciation information analysis module 3, a learner pronunciation difficulty point evaluation module 4, a learner personalized correction strategy generation module 5, a learner difficult pronunciation correction module 6, a speech model evaluation module 7 and a learning effect feedback module 8 which are connected in sequence.

[0035] The learner identification module 1 is used to identify the learner's native language and collect learner information including the learner's age, gender and English language background. Preferably, in one embodiment of the present invention, the learner's age can be 7-18 years old, the learner's native language can be Mandarin Chinese, and the learner's English language background can be zero-based, elementary, intermediate or advanced. In this way, by collecting the learner's basic information, the learner's English learning status can be better understood, providing a basis for the subsequent formulation of personalized correction strategies.

[0036] The learner pronunciation information input module 2 is used to receive the pronunciation information submitted by the learner and mark the target phonemes to be learned by the learner. Secondly, in one embodiment of the present invention, the received pronunciation information includes video information, audio information and text information, wherein the video information and audio information record the learner's pronunciation practice process, and the text information contains the voice text content. The learner pronunciation information input module 2 is also used to receive the learner's annotations on the pronunciation information, and the annotations include the pronunciation types of practice and the time intervals between different pronunciations. In this way, the learner's pronunciation information can be fully recorded, and the target phonemes that the learner wants to learn can be clearly identified through the learner's annotations, which is convenient for subsequent targeted personalized correction.

[0037] Again, the learner pronunciation information analysis module 3 is used to calculate the difference data of the acoustic features of the learner's pronunciation and the acoustic features of the target phonemes after normalization, and analyze the learner's mastery of different target phonemes based on this. Further, in one embodiment of the present invention, the learner pronunciation information analysis module 3 includes an acoustic feature extraction unit 31, a feature comparison unit 32 and a mastery calculation unit 33, wherein the acoustic feature extraction unit 31 is used to extract acoustic features from the learner's pronunciation and the target phonemes respectively, the feature comparison unit 32 is used to normalize the extracted acoustic features and calculate the difference data, and the mastery calculation unit 33 is used to calculate the learner's mastery of different target phonemes based on the difference data. By extracting and analyzing the difference in acoustic features between the learner's pronunciation and the target phonemes, the learner's mastery of different target phonemes can be quantitatively evaluated, thereby finding out the learner's pronunciation weaknesses.

[0038] Next, the learner pronunciation difficulty assessment module 4 is used to extract and analyze the pronunciation features, timing features and rule features of the learner when learning a second language, and to assess the learner's pronunciation difficulties in combination with the learner's English language background, English level and the difference data output by the pronunciation information analysis module 3. In one embodiment of the present invention, the learner pronunciation difficulty assessment module 4 includes a phoneme contrast analysis unit 41, a rule analysis unit 42 and a timing analysis unit 43, wherein the phoneme contrast analysis unit 41 is used to contrast and analyze similar phonemes in the user's native language and English, the rule analysis unit 42 is used to analyze similar rules according to preset linguistic data, and the timing analysis unit 43 is used to analyze the timing relationship between different phonemes. Preferably, the preset linguistic data includes native language linguistic data and second language learning data, wherein the native language linguistic data includes the pronunciation and pronunciation rules of the native language, and the second language learning data includes the pronunciation and pronunciation rules of the second language. In this way, the learner's pronunciation difficulties can be comprehensively assessed from the three dimensions of pronunciation features, timing features and rule features, and the key factors affecting the learner's pronunciation can be more accurately found out in combination with the learner's background information.

[0039] Then, the learner personalized correction strategy generation module 5 is used to generate a personalized correction training program including the type of corrected pronunciation, the number of exercises, the frequency of exercises, the duration of exercises, and the required learning instructors according to the pronunciation difficulties. In one embodiment of the present invention, for learners with short pronunciation duration, the correction strategy can be to increase the practice duration of the phoneme and appropriately reduce the frequency of exercises; for learners with incorrect pronunciation timing, the correction strategy can be to strengthen the connection exercises of adjacent phonemes and match instructors with rich teaching experience. By formulating personalized correction programs from multiple aspects such as pronunciation content, practice methods, and instructors for the learners' pronunciation difficulties, the learners' training efficiency and learning effect can be maximized.

[0040] Next, the learner difficult pronunciation correction module 6 is used to guide the learner to learn and correct difficult phonemes according to the personalized correction training program. In one embodiment of the present invention, the execution of the personalized correction training program includes: matching an instructor according to the learner's pronunciation difficulties; performing pronunciation correction training according to the personalized correction training program; collecting pronunciation data during the training process; and evaluating the collected pronunciation data based on the speech model. Preferably, the matching of the instructor includes: ranking the teachers according to pronunciation similarity; screening according to the evaluation of the teaching effect of the teachers; performing adaptability analysis according to the current workload of the teachers; and selecting the most suitable instructor according to the above analysis results. In this way, professional and efficient pronunciation correction guidance can be provided to the learners, and the learners' practice situation can be monitored in real time, and the training program can be adjusted in time.

[0041] Finally, the speech model evaluation module 7 is used to quantitatively evaluate the pronunciation of the learner according to the preset speech model, and the learning effect feedback module 8 is used to optimize the personalized correction training program and adjust the corresponding learning instructor according to the quantitative evaluation results. In one embodiment of the present invention, the optimization of the personalized correction training program includes: analyzing the learning effect according to the speech model evaluation results; adjusting the training content and training parameters according to the learning effect; adjusting the instructor according to the learner's feedback; and updating the pronunciation difficulty points according to the latest pronunciation data. When the speech model evaluates that the pronunciation accuracy of a certain phoneme of the learner is lower than 60%, the practice intensity of the phoneme can be appropriately increased. When the accuracy of the evaluation is higher than 90% for three consecutive times, the practice intensity can be appropriately reduced, and the practice of other phonemes can be strengthened. In this way, the learning effect of the learner can be objectively evaluated, and the personalized correction program can be dynamically optimized according to the evaluation results, while taking into account the learner's subjective feelings, and finally achieving twice the result with half the effort.

[0042] The present invention also provides an English phoneme personalized correction method based on acoustic feature analysis, which is applied to any of the above-mentioned systems and comprises the following steps:

[0043] Step S1: collecting and recording the learner's native language type, age, gender, English language background and English proficiency information through the learner identification module 1;

[0044] Step S2: receiving learner pronunciation information including video information, audio information and text information through the learner pronunciation information input module 2;

[0045] Step S3: performing acoustic feature analysis on the pronunciation information through the learner pronunciation information analysis module 3 and calculating difference data with the target phoneme;

[0046] Step S4: evaluating the learner's pronunciation difficulties according to the difference data through the learner pronunciation difficulty evaluation module 4;

[0047] Step S5: generating a personalized correction training program through the learner personalized correction strategy generation module 5;

[0048] Step S6: executing the personalized correction training program through the learner's difficult pronunciation correction module 6;

[0049] Step S7: evaluating the learning effect through the speech model evaluation module 7;

[0050] Step S8: Optimizing the personalized correction training program through the learning effect feedback module 8.

[0051] Preferably, in one embodiment of the present invention, the acoustic feature analysis includes pronunciation duration analysis, phoneme frequency analysis and pronunciation timing analysis, wherein the pronunciation duration analysis is used to analyze the phoneme duration difference between the practice pronunciation and the target pronunciation, the phoneme frequency analysis is used to analyze the phoneme frequency difference between the practice pronunciation and the target pronunciation, and the pronunciation timing analysis is used to analyze the timing state difference between the practice pronunciation and the target pronunciation, and the timing state includes the previous phoneme, the next phoneme and the timing interval between phonemes. By analyzing the acoustic features of the learner's pronunciation from multiple angles and levels, the learner's pronunciation defects can be comprehensively evaluated, providing sufficient basis for the subsequent formulation of personalized correction plans.

[0052] In summary, the English phoneme personalized correction system and method based on acoustic feature analysis provided by the present invention can intelligently identify the identity of learners, dynamically collect and analyze learners' pronunciation information, accurately evaluate learners' pronunciation difficulties, automatically generate personalized correction plans, efficiently guide learners to correct pronunciation, objectively evaluate learning effects and dynamically optimize training plans, which can maximize learners' learning efficiency and correction effects, and is of great significance to improving the level of second language learning.

[0053] The English phoneme personalized correction system based on acoustic feature analysis of the present invention, by setting a learner identity recognition module 1, a learner pronunciation information input module 2, a learner pronunciation information analysis module 3, a learner pronunciation difficulty point assessment module 4, a learner personalized correction strategy generation module 5, a learner difficult pronunciation correction module 6, a speech model assessment module 7 and a learning effect feedback module 8, each module is closely coordinated and mutually conditional, forming a set of full-process, personalized English pronunciation learning correction closed loop, which is a very beneficial attempt. The English phoneme personalized correction method based on acoustic feature analysis of the present invention provides learners with a full range of refined personalized pronunciation learning services from the links of learner information collection, pronunciation information analysis, pronunciation difficulty point assessment, personalized correction, learning effect evaluation and dynamic optimization, which is a new idea and new direction for second language education. The system and its method have good practical application value, provide efficient and practical learning tools for the majority of English learners, and will greatly improve the quality of English teaching and learning effect.

[0054] 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. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. English phoneme personalized correction system based on acoustic feature analysis, characterized by: The system comprises a learner identification module, a learner pronunciation information input module, a learner pronunciation information analysis module, a learner pronunciation difficulty assessment module, a learner personalized correction strategy generation module, a learner difficult pronunciation correction module, a speech model assessment module and a learning effect feedback module, wherein the learner identification module is used to identify the learner's native language and collect learner information including the learner's age, gender and English language background, the learner pronunciation information input module is used to receive pronunciation information submitted by the learner and have the learner mark the target phonemes to be learned, the learner pronunciation information analysis module is used to calculate the difference data of the acoustic features of the learner's pronunciation and the acoustic features of the target phonemes after normalization processing and analyze the learner's mastery of different target phonemes based on this, and the learner pronunciation difficulty assessment module is used to provide The pronunciation features, timing features and rule features of learners when learning a second language are obtained and analyzed, and the learners' pronunciation difficulties are evaluated by combining the learners' English language background, English level and the difference data output by the pronunciation information analysis module. The learner personalized correction strategy generation module is used to generate a personalized correction training plan including the type of corrected pronunciation, the number of exercises, the frequency of exercises, the length of exercises and the required learning instructors according to the pronunciation difficulties. The learner difficult pronunciation correction module is used to guide the learners to learn and correct difficult phonemes according to the personalized correction training plan. The speech model evaluation module is used to quantitatively evaluate the learners' pronunciation according to a preset speech model. The learning effect feedback module is used to optimize the personalized correction training plan according to the quantitative evaluation results and adjust the corresponding learning instructors.

2. The English phoneme personalized correction system based on acoustic feature analysis according to claim 1 is characterized in that: The pronunciation information received by the learner pronunciation information input module includes video information, audio information and text information, wherein the video information and audio information record the learner's pronunciation practice process, and the text information contains voice text content. The learner pronunciation information input module is also used to receive the learner's annotations on the pronunciation information, and the annotations include the pronunciation types of practice and the time intervals between different pronunciations.

3. The English phoneme personalized correction system based on acoustic feature analysis according to claim 1 is characterized in that: The learner pronunciation information analysis module includes an acoustic feature extraction unit, a feature comparison unit and a mastery degree calculation unit, wherein the acoustic feature extraction unit is used to extract acoustic features from the learner pronunciation and the target phonemes respectively, the feature comparison unit is used to normalize the extracted acoustic features and calculate difference data, and the mastery degree calculation unit is used to calculate the learner's mastery degree of different target phonemes based on the difference data.

4. The English phoneme personalized correction system based on acoustic feature analysis according to claim 1 is characterized in that: The learner pronunciation difficulty assessment module includes a phoneme contrast analysis unit, a rule analysis unit and a timing analysis unit, wherein the phoneme contrast analysis unit is used to contrast and analyze similar phonemes in the user's native language and English, the rule analysis unit is used to analyze similar rules based on preset linguistic data, and the timing analysis unit is used to analyze the timing relationship between different phonemes.

5. The English phoneme personalized correction system based on acoustic feature analysis according to claim 4 is characterized in that: The preset linguistic data include native language linguistic data and second language learning data, wherein the native language linguistic data includes the pronunciation and pronunciation rules of the native language, and the second language learning data includes the pronunciation and pronunciation rules of the second language.

6. A personalized English phoneme correction method based on acoustic feature analysis, applied to the system according to any one of claims 1 to 5, characterized in that: The method comprises the following steps: collecting and recording the learner's native language type, age, gender, English language background and English proficiency information through the learner identity recognition module; receiving the learner's pronunciation information including video information, audio information and text information through the learner pronunciation information input module; performing acoustic feature analysis on the pronunciation information and calculating difference data with the target phoneme through the learner pronunciation information analysis module; evaluating the learner's pronunciation difficulty points according to the difference data through the learner pronunciation difficulty point evaluation module; generating a personalized correction training plan through the learner personalized correction strategy generation module; executing the personalized correction training plan through the learner difficult pronunciation correction module; evaluating the learning effect through the speech model evaluation module; and optimizing the personalized correction training plan through the learning effect feedback module.

7. The method for personalized correction of English phonemes based on acoustic feature analysis according to claim 6, characterized in that: The acoustic feature analysis includes pronunciation duration analysis, phoneme frequency analysis and pronunciation timing analysis, wherein the pronunciation duration analysis is used to analyze the difference in phoneme duration between the practice pronunciation and the target pronunciation, the phoneme frequency analysis is used to analyze the difference in phoneme frequency between the practice pronunciation and the target pronunciation, and the pronunciation timing analysis is used to analyze the difference in timing state between the practice pronunciation and the target pronunciation, and the timing state includes the previous phoneme, the next phoneme and the timing interval between phonemes.

8. The method for personalized correction of English phonemes based on acoustic feature analysis according to claim 6, characterized in that: The execution of the personalized correction training program includes: matching an instructor according to the learner's pronunciation difficulties; performing pronunciation correction training according to the personalized correction training program; collecting pronunciation data during the training process; and evaluating the collected pronunciation data based on the speech model.

9. The method for personalized correction of English phonemes based on acoustic feature analysis according to claim 8, characterized in that: The matching of the instructors includes: ranking the instructors according to pronunciation similarity; screening the instructors according to the evaluation of the teachers' teaching effectiveness; performing adaptability analysis according to the teachers' current workload; and selecting the most suitable instructor according to the above analysis results.

10. The method for personalized correction of English phonemes based on acoustic feature analysis according to claim 6, characterized in that: The optimization of the personalized correction training program includes: analyzing the learning effect according to the speech model evaluation results; adjusting the training content and training parameters according to the learning effect; adjusting the instructor according to the learner feedback; and updating the pronunciation difficulty points according to the latest pronunciation data.

Images