Personalized note generation method and system

Abstract

An embodiment of the present invention provides a personalized note generation method. The method includes: receiving a first number of tags input by a user in real time during the playback of a media file, and recording the time information of each tag during the playback of the media file; converting the media file into a media text in real time, and dividing the media text into a second number of content blocks based on the first number of tags; performing text analysis on each content block to obtain keywords and content summaries of each content block; after the playback of the media file is completed, providing the user with personalized multi-granular notes corresponding to each content block containing time information, keywords, and content summaries. An embodiment of the present invention also provides a personalized note generation system. An embodiment of the present invention enables users to perform personalized segmentation of course audio by annotation. The generated notes are more flexible, and the keyword and summary extraction is more accurate. The obtained personalized notes can bring better help to users.

Description

Technical Field

[0001] The present invention relates to the field of intelligent speech, and in particular to a personalized note generation method and system. Background Art

[0002] With the popularity of online audio courses, the method of generating class notes has become the focus of attention of various intelligent enterprises. Currently, the generation of notes is mostly based on speech recognition technology, which converts audio signals into text data and uses it as the key content of text notes.

[0003] Text notes are usually generated by:

[0004] Voice collection module: record the teacher's lecture voice by calling the radio;

[0005] Voice conversion module: upload the recorded voice to the server, and after receiving the data, the server converts the data into text;

[0006] Note generation module: The result after speech conversion is used as the final note result.

[0007] In the process of implementing the present invention, the inventors found that there are at least the following problems in the related art:

[0008] The purpose of notes is to record key points. The course notes currently generated are all audio content, which will make the notes too redundant.

[0009] Considering that notes are for users to read, rather than simply a string of text, the course notes should be in a single text format, mostly paragraph-style text content. However, course notes can be applied to different scenarios, such as: precise review after class, systematic review before exams, etc. Therefore, designing multiple forms of notes can better meet the personalized needs of users and enhance their practicality.

[0010] Note generation based on the complete course audio does not allow users to generate segmented notes, which makes it inconvenient for users to trace the course content at a specific stage. Different users have different learning focuses for the same course, so it should be designed in a flexible way to help different users mark the key points of the course;

[0011] Course notes exist independently from course videos. Users cannot quickly find the corresponding video content based on the note content, which affects review efficiency. Summary of the invention

[0012] In order to at least solve the problems in the prior art that the generated notes are redundant, the generated notes cannot be applied to every user, the practicality can be improved, and the notes cannot be associated with the course. In a first aspect, an embodiment of the present invention provides a personalized note generation method, comprising:

[0013] During the playback of the media file, receiving in real time a first number of marks input by a user, and recording time information of each mark corresponding to the playback of the media file;

[0014] Converting the media file into a media text in real time, and dividing the media text into a second number of content blocks based on the first number of tags;

[0015] Performing text analysis on each content block to obtain keywords and content summaries of each content block;

[0016] After the media file is played, personalized multi-granular notes corresponding to each content block including time information, keywords, and content summary are fed back to the user.

[0017] In a second aspect, an embodiment of the present invention provides a personalized note generation system, including:

[0018] A tag receiving program module is used to receive a first number of tags input by a user in real time during the playback of a media file, and to record time information of each tag corresponding to the playback of the media file;

[0019] A content division program module, configured to convert the media file into a media text in real time, and to divide the media text into a second number of content blocks based on the first number of tags;

[0020] A text analysis program module, used to perform text analysis on each content block to obtain keywords and content summaries of each content block;

[0021] The note generation program module is used to provide the user with personalized multi-granular notes corresponding to each content block, including time information, keywords, and content summaries, after the media file is played.

[0022] According to a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the personalized note generation method of any embodiment of the present invention.

[0023] In a fourth aspect, an embodiment of the present invention provides a storage medium having a computer program stored thereon, characterized in that when the program is executed by a processor, the steps of the personalized note generation method of any embodiment of the present invention are implemented.

[0024] The beneficial effects of the embodiments of the present invention are as follows: users can perform personalized segmentation of course audio by annotation. The generated notes can vary from person to person and are more flexible; in the process of keyword extraction, similar words of candidate keywords are merged through text similarity calculation - this can avoid the "clustering" of keywords that may occur when directly using the tf-idf algorithm, making the keyword extraction of the target text more accurate; in the process of text summarization, the idea of ​​information entropy is introduced to extract the summary, ensuring that the summary sentence has the greatest similarity with other sentences in the cluster in the corresponding classification cluster, thereby improving the accuracy of the summary sentence. The resulting personalized multi-granular notes can bring better help to users. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0026] Figure 1 is a flow chart of a personalized note generation method provided by an embodiment of the present invention;

[0027] Figure 2 This is an overall architecture diagram of a personalized note generation method provided by an embodiment of the present invention;

[0028] Figure 3 It is a flowchart of keyword extraction based on TF-idf and text similarity of a personalized note generation method provided by an embodiment of the present invention;

[0029] Figure 4 It is a flowchart of text summary extraction based on Doc2Vec model and K-means clustering algorithm of a personalized note generation method provided by one embodiment of the present invention;

[0030] Figure 5 is a structural schematic diagram of a personalized note generation system provided by an embodiment of the present invention;

[0031] Figure 6 A schematic structural diagram of an embodiment of an electronic device for generating personalized notes provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0033] like Figure 1 The flowchart of a personalized note generation method provided by an embodiment of the present invention includes the following steps:

[0034] S11: During the playback of the media file, receiving in real time a first number of marks input by a user, and recording time information of each mark corresponding to the playback of the media file;

[0035] S12: converting the media file into a media text in real time, and dividing the media text into a second number of content blocks based on the first number of tags;

[0036] S13: Performing text analysis on each content block to obtain keywords and content summaries of each content block;

[0037] S14: After the media file is played, personalized multi-granular notes corresponding to each content block including time information, keywords, and content summary are fed back to the user.

[0038] In this embodiment, with the development of the Internet, online courses have become an important form of learning for people. In online course learning, taking notes is a common behavior of users, but the form of manually taking notes while listening to the class is likely to affect the efficiency of listening to the class. In the process of recording, because the content to be recorded will be silently recited in the mind, in this process, the content played during the recording process will be ignored. Therefore, it is very important to be able to generate user-personalized notes for the media files being played.

[0039] For step S11, during the playback of media files, for example, media files played online can also be locally cached media files. Taking into account the breadth of applicability of the present method, as an implementation method, the media files include: course audio, course video or conference audio, conference video. For example, online course videos played online or course audio cached locally can also be used to record meeting notes in consideration of wider applications. For example, important content will be played from time to time in conference audio or video, and it is also important to be able to generate personalized notes for the meeting. This method can be applied to courses and meetings, and the following takes local course videos as an example. Figure 2 As shown, this is the structure diagram generated by the notes.

[0040] When a user finds something that needs to be recorded during the playback of a local course video, he or she will mark it in the course video. Of course, there are many knowledge points in the course content, so the user will enter multiple marks. As an implementation method, the multiple marks entered by the user include: a click operation or a voice input operation.

[0041] For example, a user plays a course video on a pad using a player equipped with this method. When the key point of the course is played, the user can click on the interface for the first time. After a period of time, the key point is played, and the user can click on the decoder for the second time. When the course video plays the next key point that the user thinks is useful to him, the above click operation can be continued, and the corresponding video segment can be divided by two clicks. Or a more explicit click method, the interface provides buttons for "start recording" and "end recording", and the user can mark the course video by directly clicking.

[0042] If the pad has a voice interaction function, when the course video plays to the key points that the user thinks are useful to him, he can also input "record" by voice. When the key part is played, the user can input "stop recording" by voice. In this way, the course video can also be marked, and the corresponding video segment can be divided by two voices. Regardless of the above marking method, the time information of each mark in the course video playback will be recorded, and the information that the user wants to mark can be recorded instead of the entire video.

[0043] For step S12, after the course starts, the sound pickup is turned on, and the course audio content is converted into text based on speech recognition technology to obtain the course text. The course text is divided into multiple content blocks using the marks in step S11. For example, in the grammar course playback, the user feels that he needs to learn more about the content of "predicate" and "object". The content of "predicate" and "object" is marked in the course video. For example, the start time of "predicate" marked in the course video is 10:25, and the time of the next mark is 15:30. The start time of "object" marked in the course video is 17:18, and the time of the next mark is 22:14. In this way, the "predicate" content block of 10:25-15:30 and the "object" content block of 17:18-22:14 are obtained.

[0044] In step S13, the "predicate" content block and the "object" content block are analyzed separately to obtain the keywords and content summary of each content block;

[0045] As an implementation manner, performing text analysis on each content block to obtain keywords and content summaries of each content block includes:

[0046] Extracting the corresponding keywords from each content block based on the word frequency-inverse text frequency index and the word similarity;

[0047] The content blocks are processed based on a document vectorization model and clustering to obtain content summaries corresponding to the content blocks.

[0048] In this embodiment, keywords can be extracted based on a text keyword extraction method based on TF-idf (term frequency–inverse document frequency) and word similarity.

[0049] For example, the text content of the content block of the first paragraph is "The concept of predicate in language grammar is to indicate what the subject is like, what properties it has, what state it is in, etc. It is used to state the subject. Commonly used verbs, verb phrases, adjectives, adjective phrases, nouns, noun phrases, and subject-predicate phrases act as predicates. In English, verbs are mainly divided into two categories according to their role and function, one is predicate verbs, and the other is non-predicate verbs.

[0050] In traditional Chinese grammar, a predicate is one of the two main parts of a sentence. The other is the subject, which is modified by the predicate. The predicate provides information about the subject, such as what the subject does or is.

[0051] Through the above methods, such as Figure 3 As shown, a group of candidate words are determined by word frequency, and the "predicate" with the highest similarity is determined to be the keyword.

[0052] like Figure 4 As shown in the figure, the text summary extraction method of the Doc2Vec (document vectorization model) model and the K-means (k-means clustering algorithm) clustering algorithm can be used to generate summary texts for each module. The text of the content block is converted into a sentence vector representation, the number of summary sentences required is determined, and then K-means clustering is performed to cluster similar sentences. Finally, information entropy is used to extract key sentences, and the content summary is determined to be "stating the nature and state of the subject, and the predicate modifies the subject to indicate what the subject does." Based on the same method, the keywords and summary text of the content block of the next paragraph can be determined.

[0053] For step S14, after the course is over, the multi-granular notes containing the time information: 10:25-15:30, the keyword "predicate", the content summary "states the nature and status of the subject, the subject is modified by the predicate, and indicates what the subject does", and other multi-granular notes according to the above structure will be determined as user-personalized multi-granular notes and fed back to the user.

[0054] It can be seen from this implementation that users can personalize the segmentation of course audio by annotation. The generated notes can vary from person to person and are more flexible. In the keyword extraction process, similar words of candidate keywords are merged through text similarity calculation - this can avoid the "clustering" of keywords that may occur when using the tf-idf algorithm directly, making the keyword extraction of the target text more accurate. In the text summarization process, the idea of ​​information entropy is introduced to extract summary sentences to ensure that the summary sentences have the greatest similarity with other sentences in the cluster in the classification cluster to which they belong, thereby improving the accuracy of the summary sentences. The resulting personalized multi-granular notes can bring better help to users.

[0055] As an implementation mode, after feeding back to the user personalized multi-granular notes corresponding to each content block including time information, keywords, and content summaries, the method further includes:

[0056] In response to the user clicking on the time information in the personalized multi-granularity note, the media file is jumped to the corresponding playing time for playing.

[0057] In this embodiment, for example, in a web page video or a locally cached video, the video can be dragged. If it is a live video, the video will be stored during playback.

[0058] After users get personalized multi-granular notes, they can click on the time information, and then the course video will be played to the corresponding time. This allows users to review together with the notes and videos, improving the user experience.

[0059] In brief, this method comprises the following steps:

[0060] Step 1: Play the course audio file, obtain the user's input instructions, and mark the start time as N 0 ;

[0061] Step 2: Record the key points of the user's record in sequence, and the "Start Recording" key is marked as N i(i≥1) , "End Record" is recorded as M i(i≥1)； , and record its corresponding time position;

[0062] Step 3: Intercept N i -M iPartial audio: split the course audio into multiple independent audio segments according to user tags i , generate an audio set, denoted as S = {s i |s i =N i -M i};

[0063] Step 4: Perform speech recognition on the audio content elements in the audio set S to generate a text information set T = {T s |s=i};

[0064] Step 5: Use the Python Chinese word segmentation component Jieba to segment each independent text information s Perform word segmentation, sentence segmentation, and corpus processing;

[0065] Step 6: For each independent text T after word segmentation s ,Text keyword extraction method based on TF-idf and word similarity, extract T s The key word in is T s The corresponding audio content i Keyword information, such as Figure 3 As shown,

[0066] (1) For text T s Calculate the frequency of the words in and represent them by word vectors;

[0067] (2) Calculate the similarity of words based on cosine similarity. If the similarity between two words is greater than a certain threshold, they are considered similar words.

[0068] (3) Based on similar words, retain unique values ​​and accumulate the frequency of similar words to generate a new candidate word set;

[0069] (4) Extract keywords based on TF-IDF algorithm.

[0070] Step 7: For each independent text T after sentence segmentation s ,Based on the text summary extraction method of Doc2Vec model and K-means clustering algorithm, the summary text of each module is generated, such as Figure 4 shown.

[0071] (1) Doc2Vec model, which includes two neural network models: DM (Distributed Memory Model) and DBOW (Distributed Bag Of Words). This method directly references this model and directly converts sentences or paragraphs into fixed-dimensional vectors.

[0072] (2) Count the total number of sentences N in the text, and determine the number of summary sentences k required based on the total number of sentences N and the actual situation;

[0073] (3) Use the K-means clustering algorithm to perform similarity clustering on the sentence vectors generated by training (1) to obtain k clusters;

[0074] (4) Based on the idea of ​​information entropy, the entropy value of the similarity between the sentence vector within the cluster and other sentence vectors is calculated, and the sentence with the largest entropy value is selected as the final summary sentence output.

[0075] Step 8: Based on T s Extract keyword information and summary information from text to generate audio content in the course i , that is, N i -M i Interval, including notes of two granularities: keywords and summaries;

[0076] Step 9: Mark the course notes of each module in the course video progress bar according to the time stamp;

[0077] Step 10: Generate course audio with a personalized course note structure and containing multi-granular course note content.

[0078] like Figure 5 The figure is a schematic diagram of the structure of a personalized note generation system provided by one embodiment of the present invention. The system can execute the personalized note generation method described in any of the above embodiments and be configured in a terminal.

[0079] A personalized note generation system 10 provided in this embodiment includes: a tag receiving program module 11, a content division program module 12, a text analysis program module 13 and a note generation program module 14.

[0080] Among them, the tag receiving program module 11 is used to receive a first number of tags input by the user in real time during the playback of the media file, and record the time information of each tag corresponding to the playback of the media file; the content division program module 12 is used to convert the media file into a media text in real time, and divide the media text into a second number of content blocks based on the first number of tags; the text analysis program module 13 is used to perform text analysis on each content block to obtain keywords and content summaries of each content block; the note generation program module 14 is used to feedback personalized multi-granularity notes corresponding to each content block including time information, keywords, and content summaries to the user after the playback of the media file is completed.

[0081] Furthermore, the text analysis program module is used to:

[0082] Extracting the corresponding keywords from each content block based on the word frequency-inverse text frequency index and the word similarity;

[0083] The content blocks are processed based on a document vectorization model and clustering to obtain content summaries corresponding to the content blocks.

[0084] Furthermore, the system further comprises: a media-note association program module, which is used to:

[0085] In response to the user clicking on the time information in the personalized multi-granularity note, the media file is jumped to the corresponding playing time for playing.

[0086] The embodiment of the present invention further provides a non-volatile computer storage medium, the computer storage medium stores computer executable instructions, and the computer executable instructions can execute the personalized note generation method in any of the above method embodiments;

[0087] As an implementation mode, the non-volatile computer storage medium of the present invention stores computer executable instructions, and the computer executable instructions are configured as follows:

[0088] During the playback of the media file, receiving in real time a first number of marks input by a user, and recording time information of each mark corresponding to the playback of the media file;

[0089] Converting the media file into a media text in real time, and dividing the media text into a second number of content blocks based on the first number of tags;

[0090] Performing text analysis on each content block to obtain keywords and content summaries of each content block;

[0091] After the media file is played, personalized multi-granular notes corresponding to each content block including time information, keywords, and content summary are fed back to the user.

[0092] As a non-volatile computer-readable storage medium, it can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the method in the embodiment of the present invention. One or more program instructions are stored in the non-volatile computer-readable storage medium, and when executed by the processor, the personalized note generation method in any of the above method embodiments is executed.

[0093] Figure 6 FIG. 1 is a schematic diagram of the hardware structure of an electronic device of a personalized note generation method provided in another embodiment of the present application, such as Figure 6 As shown, the device includes:

[0094] One or more processors 610 and memory 620, Figure 6 A processor 610 is taken as an example. The device of the personalized note generation method may further include: an input device 630 and an output device 640.

[0095] The processor 610, the memory 620, the input device 630 and the output device 640 may be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0096] The memory 620, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the personalized note generation method in the embodiment of the present application. The processor 610 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 620, that is, realizing the personalized note generation method in the above method embodiment.

[0097] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data, etc. In addition, the memory 620 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 620 may optionally include a memory remotely arranged relative to the processor 610, and these remote memories may be connected to the mobile device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0098] The input device 630 can receive input digital or character information. The output device 640 can include a display device such as a display screen.

[0099] The one or more modules are stored in the memory 620, and when executed by the one or more processors 610, the personalized note generation method in any of the above method embodiments is executed.

[0100] The above-mentioned product can execute the method provided in the embodiment of the present application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of the present application.

[0101] The non-volatile computer-readable storage medium may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the device, etc. In addition, the non-volatile computer-readable storage medium may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the non-volatile computer-readable storage medium may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0102] An embodiment of the present invention also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the personalized note generation method of any embodiment of the present invention.

[0103] The electronic device of the embodiment of the present application exists in various forms, including but not limited to:

[0104] (1) Mobile communication equipment: This type of equipment is characterized by having mobile communication functions and its main purpose is to provide voice and data communications. This type of terminal includes: smart phones, multimedia phones, functional phones, and low-end phones.

[0105] (2) Ultra-mobile personal computer devices: These devices fall into the category of personal computers, have computing and processing capabilities, and generally also have mobile Internet access features. These terminals include: PDAs, MIDs, and UMPC devices, such as tablet computers.

[0106] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players, handheld game consoles, e-books, smart toys, and portable car navigation devices.

[0107] (4) Other electronic devices with data processing functions.

[0108] In this article, relational terms such as first and second, etc. are used only 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" and "comprise" include not only those elements, but also other elements not explicitly listed, or also include elements inherent to such processes, methods, articles or equipment. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the existence of other identical elements in the process, method, article or equipment that includes the elements.

[0109] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0110] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0111] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A personalized note generation method, comprising: During the playback of the media file, receiving in real time a first number of marks input by a user, and recording time information of each mark corresponding to the playback of the media file; Converting the media file into a media text in real time, and dividing the media text into a second number of content blocks based on the first number of tags; Performing text analysis on each content block to obtain words in each content block, judging the word frequency and similarity of the words in each content block, generating candidate words based on similar words, generating a new candidate word set based on the word frequency of the candidate words, extracting keywords from the candidate word set based on the TF-IDF algorithm, and obtaining a content summary; Generate personalized multi-granular notes corresponding to each content block; Mark the personalized multi-granular notes of each content block in the course video progress bar according to the time stamp.

2. The method according to claim 1, wherein: The text analysis of each content block to obtain keywords and content summaries of each content block includes: Extracting the corresponding keywords from each content block based on the word frequency-inverse text frequency index and the word similarity; The content blocks are processed based on a document vectorization model and clustering to obtain content summaries corresponding to the content blocks.

3. The method according to claim 1, wherein: After feeding back to the user the personalized multi-granularity notes corresponding to each content block, including time information, keywords, and content summaries, the method further includes: In response to the user clicking on the time information in the personalized multi-granularity note, the media file is jumped to the corresponding playing time for playing.

4. The method according to claim 1, wherein: The media files include: course audio, course video or conference audio, conference video.

5. The method according to claim 1, wherein: The multiple marks input by the user include: a click operation or a voice input operation.

6. A personalized note generation system, comprising: A tag receiving program module is used to receive a first number of tags input by a user in real time during the playback of a media file, and to record time information of each tag corresponding to the playback of the media file; A content division program module, configured to convert the media file into a media text in real time, and to divide the media text into a second number of content blocks based on the first number of tags; A text analysis program module, for performing text analysis on each content block, obtaining words of each content block, judging the word frequency and similarity of the words of each content block, generating candidate words based on similar words, generating a new candidate word set based on the word frequency of the candidate words, extracting keywords from the candidate word set based on the TF-IDF algorithm, and obtaining a content summary; The note generation program module is used to generate personalized multi-granular notes corresponding to each content block, and mark the personalized multi-granular notes of each content block in the course video progress bar according to the time mark.

7. The system according to claim 6, wherein: The text analysis program module is used to: Extracting the corresponding keywords from each content block based on the word frequency-inverse text frequency index and the word similarity; The content blocks are processed based on a document vectorization model and clustering to obtain content summaries corresponding to the content blocks.

8. The system according to claim 6, wherein: The system further comprises: a media-note association program module, for: In response to the user clicking on the time information in the personalized multi-granularity note, the media file is jumped to the corresponding playing time for playing.

9. The system according to claim 6, wherein: The media files include: course audio, course video or conference audio, conference video.

10. The system according to claim 6, wherein: The multiple marks input by the user include: a click operation or a voice input operation.

11. An electronic device, comprising: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the method described in any one of claims 1 to 5.

12. A storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 5 are implemented.

Images