Intelligent beauty course resource management system

Abstract

The invention provides an intelligent beauty course resource management system, which comprises an initial resource input module for acquiring initial resources and constructing a basic resource database; the intelligent resource classification and labeling module is used for carrying out resource category judgment and label generation to obtain a multi-dimensional label system, and then carrying out secondary rechecking and optimization on the multi-dimensional label system; the resource retrieval and recommendation module is used for constructing an index database, carrying out multi-dimensional search and relevancy sorting, and carrying out personalized recommendation by utilizing a collaborative filtering algorithm; the behavior tracking and dynamic optimization module is used for constructing a user behavior log and a recommendation model and carrying out resource recommendation optimization; and the permission and sharing module is used for performing multi-role definition and permission management, resource cross-school management and resource security management. By integrating, screening, labeling and dynamically updating various kinds of beauty digital resources, an efficient beauty resource management and application platform is provided, and accurate classification, multi-dimensional tagging management, personalized recommendation and visual management of the resources are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of aesthetic education teaching, and in particular to an intelligent aesthetic education course resource management system. Background Art

[0002] Aesthetic education is a combination of aesthetic teaching and aesthetic teaching. It aims to enhance people's ability to recognize, understand, appreciate and create beauty through education. Currently, aesthetic education is restricted by geographical location, school resources and other factors, which makes it difficult for aesthetic education to fully play its role. For example: 1) Lack of a unified resource integration platform: The resources of colleges and universities, aesthetic education institutions and independent creators are scattered, making it difficult for students and teachers to obtain aesthetic education resources. 2) Lack of intelligent management: The existing educational resource management platform has a single function and lacks customized search and recommendation. 3) Insufficient resource updating and sharing: There is a lack of evaluation mechanism for resource utilization and popularity, excellent content cannot be optimized in a timely manner, and the quality of teaching resources is difficult to continuously improve. Summary of the Invention

[0003] In order to overcome the shortcomings of the existing technology, the purpose of the present invention is to provide an intelligent aesthetic education course resource management system, which provides an efficient aesthetic education resource management and application platform by integrating, screening, labeling and dynamically updating various aesthetic education digital resources, and realizes accurate classification of resources, multi-dimensional labeling management, personalized recommendation and visual management.

[0004] To achieve the above objectives, the present invention provides the following solution: an intelligent aesthetic education curriculum resource management system, comprising:

[0005] The initial resource entry module is used to obtain initial resources using the designed automatic collection channels and manual collection channels, build a basic resource database, and generate a resource ID for each resource for basic attribute registration;

[0006] The resource intelligent classification and labeling module is used to determine the categories and generate labels for text resources, image resources, and video resources based on the basic resource database to obtain a multi-dimensional labeling system, and then conduct a secondary review and optimization of the multi-dimensional labeling system;

[0007] The resource retrieval and recommendation module is used to build an index library based on the basic resource database, perform multi-dimensional search and relevance sorting, and use collaborative filtering algorithms to design recommendation rules to make personalized recommendations for users, teachers and new users;

[0008] The behavior tracking and dynamic optimization module is used to regularly summarize and analyze the behavior data of teachers or students, obtain user behavior logs, and build a recommendation model. It combines the recommendation model and the recommendation rules to optimize resource recommendations.

[0009] The permissions and sharing module is used for multi-role definition and permissions management, cross-school resource management, and resource security management;

[0010] Among them, the initial resource entry module, the resource intelligent classification and labeling module, the resource retrieval and recommendation module, the behavior tracking and dynamic optimization module and the authority and sharing module are interconnected.

[0011] Optionally, the initial resource entry module includes:

[0012] The resource collection unit is used to define the data sources and collection scope to be collected. It uses standardized interfaces and interface documents as well as crawler technology and anti-crawling mechanisms to obtain raw resources from the defined data sources. It also sets scheduled tasks to dynamically supplement the latest resources using real-time crawling and regular iterative update mechanisms. It then cleans and structures the raw resources to obtain basic resource information, completing the design of the automatic collection channel.

[0013] The resource entry unit is used to develop a teacher upload interface on the front end to upload files and fill in attributes. It also designs a resource management backend to manage file upload records, track process status, and manage editing permissions. It then performs format verification and duplicate detection based on the uploaded files, completing the design of the manual entry channel.

[0014] A resource integration unit is used to select a distributed database and a cloud database to store the initial resources and complete the construction of a basic resource database;

[0015] An information extraction unit is used to perform text information extraction, semantic analysis, style judgment and knowledge graph association based on the basic resource database to obtain a key information set, and then store the key information set in the basic resource database;

[0016] The database unit is used to define a resource ID and add basic attributes for each resource based on the basic resource database, and introduce a relational database to store metadata and an object storage system to store unstructured data; wherein the basic attributes include resource type, subject, scope of application and copyright information.

[0017] Optionally, the information extraction unit includes:

[0018] The text information extraction subunit is used to automatically extract text information from price inquiries or images using OCR technology, and to connect the text information with metadata to supplement the description and tag attributes of the resource;

[0019] The semantic analysis subunit is used to perform word segmentation, key term extraction, and topic definition on text resources using natural language processing technology;

[0020] The style judgment subunit is used to use sentiment analysis technology to judge the style of descriptive text and add style labels;

[0021] The graph association subunit is used to use knowledge graph technology to construct an association network between the basic resource database and existing resources.

[0022] Optionally, the resource intelligent classification and labeling module includes:

[0023] An intelligent classification unit is used to generate labels for text resources, image resources, and video resources based on the basic resource database, and obtain multi-dimensional labels using a classification model based on the generated labels;

[0024] The secondary review unit is used to set the confidence level of the classification model and manually review resources with a confidence level below 85%;

[0025] The multi-dimensional label system optimization unit is used to use knowledge graph technology to construct an association network between the basic resource database and existing resources, and generate association labels based on the association network, each of which corresponds to a graph index.

[0026] Optionally, the intelligent classification unit includes:

[0027] The text classification subunit is used to perform semantic classification on text resources using NLP models, classification algorithms, and sentiment analysis techniques to obtain text labels. Then, using a multi-label learning model and an existing labeling system, the text labels are associated to obtain multiple labels for the text resources.

[0028] The image classification subunit uses convolutional neural networks and ImageNet to extract artistic elements and classify styles from image resources. It also uses color extraction algorithms and texture analysis to analyze image hue and texture to obtain multiple labels for image resources.

[0029] The video classification subunit is used to perform key frame processing and image analysis on video resources, and use MFCC to analyze the audio melody characteristics in the video to obtain multiple labels for the video resources;

[0030] The label classification subunit is used to generate multi-dimensional labels including art categories, historical periods, style characteristics, teaching objectives and curriculum standards using a classification model based on the text resource multi-label, the image resource multi-label and the video resource multi-label.

[0031] Optionally, the resource retrieval and recommendation module includes:

[0032] An index library unit, configured to create a distributed index architecture, based on which full-text indexes are created for text resources, keyword indexes are created for tags and classification fields, and numerical and Boolean indexes are created for resource attributes;

[0033] The search function unit is used to implement multi-dimensional screening and multi-condition combination query, and use the BM25 algorithm to sort the search results by relevance;

[0034] The recommendation engine unit is used to calculate the similarity between users and resources based on user behavior data and utilize collaborative filtering algorithms. It then designs recommendation rules for teachers, students, and new users to provide personalized recommendations.

[0035] The visualization unit is used to define the retrieval interface including the search box, filter bar and result display, as well as the recommendation interface including the personalized recommendation list and recommendation path guidance, and display the sorting logic of the retrieval results and recommendation results.

[0036] Optionally, the recommendation engine unit includes:

[0037] The data collection subunit is used to automatically record the user's use of resources, clean and normalize the use behavior, obtain user behavior data, and then use the NoSQL database to store the user behavior data;

[0038] A similarity calculation subunit, configured to calculate the similarity between users and the similarity between resources using user-based collaborative filtering and item-based collaborative filtering, respectively;

[0039] The recommendation logic sub-unit is used to recommend similar theme resources and complementary resources based on the theme of the teacher's teaching course, complete the design of teacher recommendation rules, recommend the same category style, expanded style and cross-domain similar style based on the student learning path, complete the design of student recommendation rules, recommend popular resources based on the resource heat index and the relevance of the system default tag, and complete the design of new user recommendation rules; wherein, the resource heat index includes collection, rating and download count.

[0040] Optionally, the behavior tracking and dynamic optimization module includes:

[0041] A log recording unit, configured to extract and record the behavior type, behavior timestamp, click source page, user information, resource information, and user device based on the user behavior data to obtain a user behavior log, and then store the user behavior log in the basic resource data;

[0042] a behavior analysis and modeling unit, configured to periodically calculate resource popularity and resource utilization indicators, select a deep learning model, train the selected deep learning model using the user behavior data, the resource popularity, and the resource utilization indicators to obtain a recommendation model, and optimize resource recommendations using the recommendation model and the recommendation rules; wherein the resource utilization indicators include data indicators and categorized resource utilization indicators;

[0043] The resource library dynamic update unit is used to design high-quality resource screening rules to regularly generate a top resource list, and define resource elimination rules to regularly filter the basic resource data; wherein, the high-quality resource screening rules include a popularity rating threshold, teaching feedback, and cumulative number of collections, and the resource elimination rules include a resource usage rate threshold and teaching feedback.

[0044] Optionally, the resource heat expression is:

[0045] U1=ω1·a+ω2·b+ω3·c+ω4·d

[0046] Among them, U1 is the popularity score, ω1, ω2, ω3, and ω4 are all behavioral weights, a is the number of views, b is the number of collections, c is the number of downloads, and d is the evaluation weight;

[0047] The expression of the data indicator is:

[0048]

[0049] Among them, U2 is the resource utilization rate, c is the total number of resource accesses, c0 is the number of invalid resource accesses, and t is the resource availability time.

[0050] Optionally, the permission and sharing module includes:

[0051] A multi-role management unit is used to define system roles, role permissions, and multi-level resource reading permissions; wherein the multi-level reading permissions include public resource reading permissions that allow all users to access and download, editable resource reading permissions that allow teachers to download, and paid resource reading permissions;

[0052] An inter-school sharing unit is used to select inter-school resources from the basic resource database, define sharing conditions for the inter-school resources, sign a cooperation agreement based on the sharing conditions, and authorize resource sharing;

[0053] The resource security unit is used to set the access key of the basic resource database and add watermarks and copyright ownership when previewing and downloading resources.

[0054] The present invention provides an intelligent aesthetic education course resource management system, which has the following technical effects:

[0055] 1. Multi-dimensional Resource Classification and Intelligent Labeling: 1) By employing machine learning algorithms such as text clustering and image recognition, we enable intelligent classification and labeling of all resources (e.g., art form, historical period, style, and grade level). 2) We support secondary manual review and adjustment to ensure the accuracy and scientific nature of the classification system. 3) By establishing a multi-dimensional labeling system (including art categories, technical themes, teaching objectives, and curriculum standards), resources can be simultaneously classified into multiple labeling dimensions.

[0056] 2. Personalized, precise recommendations: By building an index library, enabling multi-dimensional search and relevance sorting, and using algorithms like collaborative filtering, precise recommendations are achieved. For example, collaborative filtering algorithms can recommend resources similar to or complementary to a teacher's frequently taught topics; for students, they can recommend art materials of similar style or higher difficulty based on their frequently viewed music.

[0057] 3. Real-time Updates and Dynamic Optimization: 1) Record resource clicks, favorites, downloads, and reviews to generate resource popularity analysis, providing data support for the system's automatic recommendations. 2) Regularly evaluate resource usage and teaching feedback, automatically screening high-quality content and highlighting resources that need to be added or eliminated, ensuring continuous optimization and updating of the resource library.

[0058] 4. Flexible Permissions: 1) Provides two-level permission management within and across campuses. Intra-campus teaching and research groups can share resources, and inter-campus alliances can share high-quality resources according to cooperation agreements. 2) Supports multi-level reading permissions for resources, including public resources, read-only reference, downloadable and editable, and paid resources, ensuring flexibility in copyright and resource management.

[0059] 5. Human-computer collaborative management: Teachers can not only enjoy the intelligent recommendations generated by the system, but also adjust and supplement resources on their own to achieve flexible teaching.

[0060] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] 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. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0062] Figure 1 A schematic diagram of the system architecture provided by an embodiment of the present invention;

[0063] Figure 2A schematic diagram of the resource classification and labeling process provided by an embodiment of the present invention;

[0064] Figure 3 A schematic diagram of the personalized recommendation process provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0066] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0067] like Figure 1 As shown, the present invention provides an intelligent aesthetic education course resource management system, including an interconnected initial resource entry module, a resource intelligent classification and labeling module, a resource retrieval and recommendation module, a behavior tracking and dynamic optimization module, and a permission and sharing module.

[0068] 1. Initial resource entry module

[0069] It is used to obtain initial resources using the designed automatic collection channel and manual collection channel, and build a basic resource database, generating a resource ID for each resource for basic attribute registration. The initial resource entry module includes:

[0070] 1.1 Resource Collection Unit

[0071] Define the data sources and scope to be collected. Target data sources include open digital materials from art institutions such as art galleries, music theaters, and dance academies, copyright-authorized resource databases, open educational resource platforms (such as Khan Academy), and public aesthetic education databases from major universities. Resource types include art works, music videos, dance teaching materials, multimedia courseware, and other content relevant to aesthetic education.

[0072] Utilize standardized interfaces and interface documents as well as crawler technology and anti-crawling mechanisms to obtain original resources from defined data sources, set timed tasks, and utilize real-time crawling and regular iterative update mechanisms to dynamically supplement the latest resources. Then, perform data cleaning on the original resources, remove invalid data such as irrelevant resources and duplicate content, fill in resource metadata such as marking missing categories or age information, and structure them to obtain basic resource information, thus completing the design of the automatic collection channel.

[0073] Aesthetic education curriculum resources include:

[0074] Art theory knowledge: theories of art forms such as painting, music, dance, drama, film and television, such as color theory in painting and music theory in music.

[0075] Appreciation of Artworks: Analysis of classic artworks from ancient and modern times, both Chinese and foreign, and their background, style, and techniques to help understand the connotations of the works.

[0076] Humanistic Literacy / Spirit: Knowledge in fields such as history, literature, philosophy, and art. For example, students should be aware of the rise and fall of important civilizations in history, classic literary works, and major schools of philosophical thought. This includes improving students' verbal expression, reading comprehension, critical thinking, and aesthetic appreciation. This includes being able to clearly and accurately express their views, deeply interpret texts, rationally analyze and judge various phenomena, and appreciate and evaluate the beauty of works of art. It also includes possessing good moral qualities and values, such as respecting others, caring for life, being honest and trustworthy, and having a sense of social responsibility. They should adhere to ethical norms and pursue moral perfection in their behavior. The humanistic spirit includes concern for the fate of humanity, the pursuit of justice and fairness, and respect for and tolerance of cultural diversity. For example, students should actively participate in public welfare and work to improve human living conditions; advocate for social fairness and justice, oppose discrimination and oppression; and respect the differences among different ethnic groups and cultures, promoting cultural exchange and integration.

[0077] Chinese traditional culture and skills: philosophy, literature, art, handicrafts, architecture, festival customs, etc.

[0078] Philosophical thoughts: including Confucianism, Taoism, Mohism, Legalism and many other schools.

[0079] Literature and arts include:

[0080] Literature: It encompasses a variety of forms, including poetry, prose, and novels. From the Book of Songs and Chu Ci to Tang poetry, Song lyrics, Yuan opera, and Ming and Qing novels, countless classic works and outstanding writers have emerged.

[0081] Arts: includes music, dance, painting, calligraphy, etc. Traditional Chinese music includes repertoire played on instruments such as the guqin and guzheng, dance includes classical dance, painting includes landscape paintings and figure paintings, and calligraphy includes various scripts such as seal script, official script, regular script, running script, and cursive script.

[0082] Languages ​​include:

[0083] Chinese character culture: From ancient oracle bone inscriptions and bronze inscriptions to seal script, clerical script, regular script, cursive script, running script, etc., the evolution of Chinese characters carries rich historical and cultural information.

[0084] Language and Culture: Chinese is the main language, and there are also many minority languages ​​and local dialects, which together constitute a rich and colorful language system.

[0085] Traditional skills include

[0086] Medicine: Traditional Chinese Medicine and Chinese Herbal Medicine have a complete theoretical and practical system, with unique diagnostic methods such as inspection, auscultation, inquiry and palpation, as well as treatment methods such as acupuncture, moxibustion and Chinese herbal prescriptions.

[0087] Handicrafts: such as ceramics, glass, lacquerware, silk, paper cutting, embroidery, wood carving, jade carving, cloisonné, etc., with exquisite craftsmanship and extremely high artistic value.

[0088] Architecture and Construction: Ancient Chinese buildings such as palaces, temples, and gardens are unique in their architectural style, layout, and construction techniques, reflecting the wisdom and aesthetic taste of the ancients.

[0089] Traditional customs include:

[0090] Traditional festivals: such as the Spring Festival, Lantern Festival, Qingming Festival, Dragon Boat Festival, Mid-Autumn Festival, etc. Each festival has specific customs and cultural connotations, carrying people's emotions and wishes.

[0091] Folk customs: including folk beliefs, folk witchcraft, etiquette and customs, wedding customs, etc., reflect the folk lifestyle and value orientation.

[0092] Local culture: local music, dance, architecture, folk heritage, etc.

[0093] Professional ethics cases: Ethical ethics cases in various professional fields for students to match their majors.

[0094] Teaching courseware: Covering various aesthetic education courses, such as painting techniques in art classes and musical note recognition in music classes, these courses are presented through a combination of graphics, text, and animations, making the knowledge easier to understand. For example, when explaining color matching, dynamic demonstrations are used to show the effects of different color combinations.

[0095] Teaching videos: including teaching demonstration videos, such as a teacher demonstrating standard movements in a dance class; art appreciation videos, in-depth analysis of the connotations of classic works, such as explaining the creative background and artistic value of Leonardo da Vinci's "Mona Lisa"

[0096] Interactive resources: Develop interactive tools such as online painting and music creation to allow students to practice their creations; set up art knowledge Q&A and discussion areas to stimulate students' thinking and communication, such as conducting an online discussion on the artistic techniques of a certain movie.

[0097] 1.2 Resource Entry Unit

[0098] It is used to develop the teacher upload interface on the front end to upload files and fill in attributes, and to design the resource management background to manage file upload records, track process status and manage editing permissions. It then performs format verification and duplicate detection based on the uploaded files to complete the design of the manual entry channel.

[0099] File upload: Teachers can upload files in various formats such as pictures, videos (MP4 format), PPT, PDF, etc.

[0100] Attribute filling: guide teachers to fill in the attributes of resources according to the template, such as title, course objectives, and applicable scope of the learning stage;

[0101] Upload record management: teachers can view the uploaded history;

[0102] Process status: displays whether the resource has been successfully imported into the database and whether the review has been completed;

[0103] Editing permissions: Allow teachers to edit or delete uploaded content.

[0104] 1.3 Resource Integration Unit

[0105] It is used to select distributed databases and cloud databases to store the initial resources and complete the construction of the basic resource database.

[0106] 1.4 Information Extraction Unit

[0107] It is used to perform text information extraction, semantic analysis, style judgment and knowledge graph association based on the basic resource database to obtain a key information set, and then store the key information set in the basic resource database.

[0108] The information extraction unit includes:

[0109] 1.4.1 Text Information Extraction Subunit

[0110] It is used to automatically extract text information from price inquiries or images using OCR technology, and connect the text information with metadata to supplement the description and tag attributes of the resource.

[0111] 1.4.2 Semantic Analysis Subunit

[0112] It is used to use natural language processing technology to perform word segmentation, key term extraction (artistic genres, historical periods) and topic definition on text resources such as lesson plans or texts.

[0113] 1.4.3 Style Judgment Subunit

[0114] It is used to use sentiment analysis technology to judge the style of descriptive text, such as beautiful, warm, soft, etc., and add style labels.

[0115] 1.4.4 Map-associated subunits

[0116] It is used to use knowledge graph technology to build an association network between the basic resource database and existing resources, for example, "Picasso paintings" are automatically associated with the "Cubism" style label.

[0117] 1.5 Database Unit

[0118] It is used to define resource IDs and add basic attributes for each resource based on the basic resource database, introduce relational databases for metadata storage, and introduce object storage systems for unstructured data storage; wherein, the basic attributes include resource type, subject, scope of application and copyright information.

[0119] Define the resource ID format: for example: `R20250409-001`, which includes the date and serial number to ensure that data resources collected from multiple sources will not be overwritten due to naming conflicts.

[0120] Basic attribute fields include but are not limited to: types such as video, picture, text, courseware, topics such as "modern dance teaching", applicable scope such as middle school / university art courses, and copyright information.

[0121] 2. Resource intelligent classification and labeling module

[0122] like Figure 2 As shown, it is used to perform category judgment and label generation for text resources, image resources and video resources based on the basic resource database to obtain a multi-dimensional label system, and then conduct a secondary review and optimization of the multi-dimensional label system. The resource intelligent classification and labeling module includes:

[0123] 2.1 Intelligent Classification Unit

[0124] It is used to generate labels for text resources, image resources and video resources based on the basic resource database, and obtain multi-dimensional labels using a classification model based on the generated labels.

[0125] Smart classification units include:

[0126] 2.1.1 Text Classification Subunit

[0127] It is used to use NLP models, classification algorithms and sentiment analysis technology to perform semantic classification on text resources to obtain text labels, and then use multi-label learning models and existing label systems to associate the text labels to obtain multiple labels for text resources.

[0128] Semantic classification: Use classification algorithms, such as support vector machines and BERT classification models, to classify text into fixed categories, such as artistic genres and style characteristics, and calculate the weight of keyword appearances for tag weight sorting.

[0129] Sentiment analysis: Analyze the intention or emotion of the content text, such as whether the text description conveys artistic characteristics such as "warmth" and "freedom".

[0130] Multi-label generation: Through multi-label learning models, such as ML-KNN, resources are labeled with multiple related labels, such as "Impressionism, prominent colors, 1875", associated with the existing label system, and dynamically supplemented with topic labels.

[0131] 2.1.2 Image Classification Subunit

[0132] It is used to extract artistic elements and classify styles of image resources using convolutional neural networks and ImageNet, and to analyze image tone and texture using color extraction algorithms and texture analysis to obtain multiple labels for image resources.

[0133] Artistic element extraction: "Portrait", "Natural Landscape", "Abstract Art", etc. Style classification: "Oil Painting", "Sketch", "Sculpture".

[0134] 2.1.3 Video Classification Subunit

[0135] It is used to perform key frame processing and image analysis on video resources, use MFCC to analyze the audio melody characteristics in the video, and obtain multiple labels for video resources.

[0136] Video frame segmentation and image analysis: The video resource is broken down into frames (with a frame interval set to 1 second), and the key frames are processed using the same image classification technology. Core classification information, such as "dance type" and "movement details," is extracted from the frame content.

[0137] Audio feature recognition: Extract the audio part of the video, use MFCC (Mel-Frequency Cepstral Coefficient) to analyze the melody characteristics, and use the classification model to identify the music type and emotional characteristics corresponding to the audio.

[0138] 2.1.4 Label Classification Subunit

[0139] It is used to generate multi-dimensional labels including art categories, historical periods, style characteristics, teaching objectives and curriculum standards using a classification model based on the text resource multi-labels, the image resource multi-labels and the video resource multi-labels.

[0140] Art Categories: Fine Arts, Music, Dance, Drama, etc. Historical Periods: Classical, Modern, Contemporary, etc. Stylistic Characteristics: Abstract, Romanticism, Realism, etc. Teaching Objectives: Inspire thinking, train skills, and enhance cultural cognition. Multi-dimensional Curriculum Standards: Elementary School Art Curriculum, Middle and High School Aesthetic Education, etc.

[0141] 2.2 Secondary Review Unit

[0142] Used to set the confidence level of the classification model and manually review resources with a confidence level below 85% to ensure the standardization and accuracy of labels.

[0143] 2.3 Multi-dimensional labeling system optimization unit

[0144] It is used to use knowledge graph technology to build an association network between the basic resource database and existing resources, and generate association tags based on the association network. Each of the association tags corresponds to a graph index, building a more friendly retrieval system to facilitate users to consult the associated content of related resources through intuitive interaction.

[0145] 3. Resource retrieval and recommendation module

[0146] like Figure 3 As shown, it is used to build an index library based on the basic resource database, perform multi-dimensional search and relevance sorting, and use collaborative filtering algorithms to design recommendation rules to make personalized recommendations for users, teachers and new users. The resource retrieval and recommendation module includes:

[0147] 3.1 Index Library Unit

[0148] It is used to create a distributed index architecture. Based on the distributed index architecture, full-text indexes are created for text resources, keyword indexes are created for tags and classification fields, and numerical and Boolean indexes are created for resource attributes.

[0149] Full-text indexing: supports fuzzy search and phonetic matching. Keyword indexing: quickly supports filtering based on exact matching. Numeric and Boolean indexing: supports range queries.

[0150] 3.2 Search Function Unit

[0151] It is used to implement multi-dimensional screening and multi-condition combination queries, and uses the BM25 algorithm to sort the search results by relevance.

[0152] Multi-dimensional filtering: Supports fine-grained filtering based on category, difficulty, applicable grade, and resource type. Combined with Boolean query logic, it supports multi-condition combination queries. Resource type: For example, text / image / video. Multi-condition combination queries: For example, "art + high difficulty + middle school and high school resources."

[0153] 3.3 Recommendation Engine Unit

[0154] Based on user behavior data, the collaborative filtering algorithm is used to calculate the similarity between users and resources, and then design recommendation rules for teachers, students, and new users to complete personalized recommendations. The recommendation engine unit includes:

[0155] 3.3.1 Data Collection Subunit

[0156] It is used to automatically record user's resource usage behavior, clean and normalize the usage behavior, obtain user behavior data, and then use the NoSQL database to store the user behavior data and establish a relationship mapping in the format of `user ID-resource ID`.

[0157] Cleaning: Removes noise data, such as short visits and empty clicks. Normalization: Normalizes behavioral data such as browsing time and ratings to balance the weights of different behaviors.

[0158] Usage behavior includes: browsing time, collection or downloading behavior, frequency of use in course teaching scenarios, search keyword habits, number of likes or comments, etc.

[0159] 3.3.2 Similarity Calculation Subunit

[0160] It is used to calculate the similarity between users and the similarity between resources using user-based collaborative filtering and item-based collaborative filtering, respectively.

[0161] User-Based Collaborative Filtering (CF): Calculates similarity between users. For example, if user A and user B both browsed resources on "Impressionist Painting" and "Baroque Music," they are considered to have similar interests. This similarity is calculated using cosine similarity or the Pearson correlation coefficient. Resources are recommended based on similar users, for example, recommending resources of interest to user B.

[0162] Item-Based Collaborative Filtering (Item-Based CF): Calculates the similarity between resources. For example, if resource A (teaching video: famous painting appreciation) and resource B (text: artist biography) both belong to the "Impressionism" category and have similar styles, resources similar to those viewed by the user will be recommended first.

[0163] 3.3.3 Recommended Logical Subunits

[0164] Teacher recommendation rules: Based on the theme of the teacher's course, we generate a teacher "interest profile" and recommend resources on similar themes and complementary resources. Resources on similar themes include: "Impressionist Famous Paintings Lecture" → "Course Video": Monet's Light and Shadow. Complementary resources include: corresponding homework materials, teaching reflection cases, etc.

[0165] Student recommendation rules: Based on the student's learning path, we recommend similar styles, expanded styles, and cross-disciplinary similar styles. Expanded styles: For example, Impressionism Basics → Advanced Painting Course. Cross-disciplinary similar styles: For example, Impressionist Painting → Impressionist Music.

[0166] Recommendation rules for new users: Recommend popular resources based on resource popularity metrics and system default tag relevance. Resource popularity metrics include favorites, ratings, and download counts. Popular resources include short videos on popular teaching topics.

[0167] 3.4 Visualization Unit

[0168] 1) Define the search interface:

[0169] Search box: supports dynamic keyword prompts in the form of drop-down association.

[0170] Filter bar: Intuitively displays optional conditions in the form of multiple-choice buttons and sliders, such as difficulty star rating and applicable scenarios.

[0171] Result display: Display search results in the form of cards, including resource thumbnails, titles, descriptions, and tags.

[0172] 2) Define the recommendation interface:

[0173] Personalized recommendation list: modular display such as "Guess you like" and "Resources used by other teachers".

[0174] Recommended path guidance: Enhance system navigation capabilities by visually displaying learning routes, such as "beginner video → intermediate text".

[0175] 3) Result sorting and explanation: Both search and recommendation results clearly display the sorting logic, such as "based on similarity" or "user rating", to improve the transparency of recommendations.

[0176] 4. Behavior tracking and dynamic optimization module

[0177] like Figure 3 As shown, it is used to regularly summarize and analyze the behavior data of teachers or students, obtain user behavior logs, and build a recommendation model, combining the recommendation model and the recommendation rules to optimize resource recommendations.

[0178] The behavior tracking and dynamic optimization module includes:

[0179] 4.1 Logging Unit

[0180] It is used to extract and record the behavior type, behavior timestamp, click source page, user information, resource information and user device based on the user behavior data, obtain the user behavior log, and then store the user behavior log in the basic resource data.

[0181] Behavior types: browsing (time and frequency of accessing a resource), collecting (adding a resource to the user's favorites), downloading (whether the user downloads the resource for local use), evaluation (including rating items and textual feedback, such as the length of the evaluation content and sentiment analysis results), and resource dwell time (the actual length of time the user stays on the browsing interface).

[0182] User information: such as user ID, identity: teacher / student.

[0183] Resource information: resource ID, resource type, and tags.

[0184] User device: device model, browser, operating system, etc., for additional behavioral analysis.

[0185] 4.2 Behavior Analysis and Modeling Unit

[0186] The system is used to regularly calculate (and update) resource popularity and resource utilization indicators, select a deep learning model, use the user behavior data, the resource popularity, and the resource utilization indicators to train the selected deep learning model to obtain a recommendation model, and use the recommendation model and the recommendation rules to optimize resource recommendations. The resource utilization indicators include data indicators and category-specific resource utilization indicators.

[0187] The expression of the resource heat is:

[0188] U1=ω1·a+ω2·b+ω3·c+ω4·d

[0189] Among them, U1 is the popularity score, ω1, ω2, ω3, and ω4 are all behavioral weights, a is the number of views, b is the number of collections, c is the number of downloads, and d is the evaluation weight;

[0190] The expression of the data indicator is:

[0191]

[0192] Among them, U2 is the resource utilization rate, c is the total number of resource accesses, c0 is the number of invalid resource accesses, and t is the resource availability time.

[0193] 4.3 Resource Library Dynamic Update Unit

[0194] Used to design high-quality resource screening rules to regularly generate a top resource list, and define resource elimination rules to regularly filter the basic resource data; wherein, the high-quality resource screening rules include popularity score thresholds, teaching feedback, and cumulative collection times, and the resource elimination rules include resource usage rate thresholds and teaching feedback.

[0195] 1) High-quality resource screening rules:

[0196] Heat score threshold: Use heat score exceeding the set threshold, such as the top 20% of resources;

[0197] Teaching feedback, including average ratings, such as the average of 5-star ratings. Comment analysis: sentiment analysis of the comments to extract positive or negative keywords. For example, if the average rating is > 4, the proportion of positive sentiment is > 70%;

[0198] The cumulative number of collections is high, such as more than five collections.

[0199] 2) Resource elimination rules

[0200] The usage rate has been below the threshold for a long time, such as less than 5 searches in the past year, and teaching feedback has been marked as negative many times, such as the average rating is less than 2 stars.

[0201] 5. Permissions and Sharing Module

[0202] It is used for multi-role definition and permission management, cross-school resource management and resource security management. The permission and sharing module includes:

[0203] 5.1 Multi-role management unit

[0204] Used to define system roles, role permissions, and multi-level resource reading permissions; wherein, the multi-level reading permissions include:

[0205] Public Resource Reading Permissions: Everyone (all user roles) can access and download resources, such as teaching examples and open-source art materials. Read-Only Resources: Content can be browsed but not downloaded or edited, suitable for copyrighted resources. Download permissions require application and can only be downloaded after the administrator reviews and authorizes.

[0206] Editable resource reading permission: allows teachers to download, modify and re-upload versions, and the version history is stored in the resource log.

[0207] Reading rights for paid resources: You need to pay a fee to read or download. Different schools can set differentiated charging standards according to the cooperation agreement.

[0208] 5.2 Cross-school shared units

[0209] It is used to select cross-school resources in the basic resource database, define the sharing conditions of the cross-school resources, sign a cooperation agreement based on the sharing conditions, and perform resource sharing authorization.

[0210] 5.3 Resource Security Unit

[0211] It is used to set the access key of the basic resource database and add watermarks and copyright attribution when previewing and downloading resources to prevent illegal dissemination.

[0212] In summary, this invention can address the problems of scattered and fragmented resources, as well as the lack of systematic management. It intelligently screens, categorizes, and integrates resources from multiple sources, including global, national, local, and institutional sources. This enriches teaching content and promptly reflects the latest developments in the art field. It also improves resource utilization efficiency, enabling both teachers and students to easily access and utilize resources, and enabling collaborative resource sharing with institutions, art organizations, and society.

[0213] Therefore, the present invention provides an intelligent aesthetic education course resource management system, which integrates, screens, labels and dynamically updates various aesthetic education digital resources to provide an efficient aesthetic education resource management and application platform, thereby realizing accurate classification of resources, multi-dimensional labeling management, personalized recommendation and visual management.

[0214] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0215] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The above examples are only intended to help understand the method and core concept of the present invention. At the same time, those skilled in the art will find that the specific implementation methods and application scopes may vary based on the concept of the present invention. In summary, the contents of this specification should not be construed as limiting the present invention.

Claims

1. An intelligent aesthetic education curriculum resource management system, characterized by: include: The initial resource entry module is used to obtain initial resources using the designed automatic collection channels and manual collection channels, build a basic resource database, and generate a resource ID for each resource for basic attribute registration; The resource intelligent classification and labeling module is used to determine the categories and generate labels for text resources, image resources, and video resources based on the basic resource database to obtain a multi-dimensional labeling system, and then conduct a secondary review and optimization of the multi-dimensional labeling system; The resource retrieval and recommendation module is used to build an index library based on the basic resource database, perform multi-dimensional search and relevance sorting, and use collaborative filtering algorithms to design recommendation rules to make personalized recommendations for users, teachers and new users; The behavior tracking and dynamic optimization module is used to regularly summarize and analyze the behavior data of teachers or students, obtain user behavior logs, and build a recommendation model. It combines the recommendation model and the recommendation rules to optimize resource recommendations. The permissions and sharing module is used for multi-role definition and permissions management, cross-school resource management, and resource security management; Among them, the initial resource entry module, the resource intelligent classification and labeling module, the resource retrieval and recommendation module, the behavior tracking and dynamic optimization module and the authority and sharing module are interconnected.

2. An intelligent aesthetic education curriculum resource management system according to claim 1, characterized in that: The initial resource entry module includes: The resource collection unit is used to define the data sources and collection scope to be collected. It uses standardized interfaces and interface documents as well as crawler technology and anti-crawling mechanisms to obtain raw resources from the defined data sources. It also sets scheduled tasks to dynamically supplement the latest resources using real-time crawling and regular iterative update mechanisms. It then cleans and structures the raw resources to obtain basic resource information, completing the design of the automatic collection channel. The resource entry unit is used to develop a teacher upload interface on the front end to upload files and fill in attributes. It also designs a resource management backend to manage file upload records, track process status, and manage editing permissions. It then performs format verification and duplicate detection based on the uploaded files, completing the design of the manual entry channel. A resource integration unit is used to select a distributed database and a cloud database to store the initial resources and complete the construction of a basic resource database; An information extraction unit is used to perform text information extraction, semantic analysis, style judgment and knowledge graph association based on the basic resource database to obtain a key information set, and then store the key information set in the basic resource database; The database unit is used to define a resource ID and add basic attributes for each resource based on the basic resource database, and introduce a relational database to store metadata and an object storage system to store unstructured data; wherein the basic attributes include resource type, subject, scope of application and copyright information.

3. An intelligent aesthetic education curriculum resource management system according to claim 2, characterized in that: The information extraction unit includes: The text information extraction subunit is used to automatically extract text information from price inquiries or images using OCR technology, and to connect the text information with metadata to supplement the description and tag attributes of the resource; The semantic analysis subunit is used to perform word segmentation, key term extraction, and topic definition on text resources using natural language processing technology; The style judgment subunit is used to use sentiment analysis technology to judge the style of descriptive text and add style labels; The graph association subunit is used to use knowledge graph technology to construct an association network between the basic resource database and existing resources.

4. An intelligent aesthetic education curriculum resource management system according to claim 3, characterized in that: The resource intelligent classification and labeling module includes: An intelligent classification unit is used to generate labels for text resources, image resources, and video resources based on the basic resource database, and obtain multi-dimensional labels using a classification model based on the generated labels; The secondary review unit is used to set the confidence level of the classification model and manually review resources with a confidence level below 85%; The multi-dimensional label system optimization unit is used to use knowledge graph technology to construct an association network between the basic resource database and existing resources, and generate association labels based on the association network, each of which corresponds to a graph index.

5. An intelligent aesthetic education curriculum resource management system according to claim 4, characterized in that: The intelligent classification unit includes: The text classification subunit is used to perform semantic classification on text resources using NLP models, classification algorithms, and sentiment analysis techniques to obtain text labels. Then, using a multi-label learning model and an existing labeling system, the text labels are associated to obtain multiple labels for the text resources. The image classification subunit uses convolutional neural networks and ImageNet to extract artistic elements and classify styles from image resources. It also uses color extraction algorithms and texture analysis to analyze image hue and texture to obtain multiple labels for image resources. The video classification subunit is used to perform key frame processing and image analysis on video resources, and use MFCC to analyze the audio melody characteristics in the video to obtain multiple labels for the video resources; The label classification subunit is used to generate multi-dimensional labels including art categories, historical periods, style characteristics, teaching objectives and curriculum standards using a classification model based on the text resource multi-label, the image resource multi-label and the video resource multi-label.

6. An intelligent aesthetic education curriculum resource management system according to claim 5, characterized in that: The resource retrieval and recommendation module includes: An index library unit, configured to create a distributed index architecture, based on which full-text indexes are created for text resources, keyword indexes are created for tags and classification fields, and numerical and Boolean indexes are created for resource attributes; The search function unit is used to implement multi-dimensional screening and multi-condition combination query, and use the BM25 algorithm to sort the search results by relevance; The recommendation engine unit is used to calculate the similarity between users and resources based on user behavior data and utilize collaborative filtering algorithms. It then designs recommendation rules for teachers, students, and new users to provide personalized recommendations. The visualization unit is used to define the retrieval interface including the search box, filter bar and result display, as well as the recommendation interface including the personalized recommendation list and recommendation path guidance, and display the sorting logic of the retrieval results and recommendation results.

7. An intelligent aesthetic education curriculum resource management system according to claim 6, characterized in that: The recommendation engine unit includes: The data collection subunit is used to automatically record the user's use of resources, clean and normalize the use behavior, obtain user behavior data, and then use the NoSQL database to store the user behavior data; A similarity calculation subunit, configured to calculate the similarity between users and the similarity between resources using user-based collaborative filtering and item-based collaborative filtering, respectively; The recommendation logic sub-unit is used to recommend similar theme resources and complementary resources based on the theme of the teacher's teaching course, complete the design of teacher recommendation rules, recommend the same category style, expanded style and cross-domain similar style based on the student learning path, complete the design of student recommendation rules, recommend popular resources based on the resource heat index and the relevance of the system default tag, and complete the design of new user recommendation rules; wherein, the resource heat index includes collection, rating and download count.

8. An intelligent aesthetic education curriculum resource management system according to claim 7, characterized in that: The behavior tracking and dynamic optimization module includes: A log recording unit, configured to extract and record the behavior type, behavior timestamp, click source page, user information, resource information, and user device based on the user behavior data, to obtain a user behavior log, and then store the user behavior log in the basic resource data; a behavior analysis and modeling unit, configured to periodically calculate resource popularity and resource utilization indicators, select a deep learning model, train the selected deep learning model using the user behavior data, the resource popularity, and the resource utilization indicators to obtain a recommendation model, and optimize resource recommendations using the recommendation model and the recommendation rules; wherein the resource utilization indicators include data indicators and categorized resource utilization indicators; The resource library dynamic update unit is used to design high-quality resource screening rules to regularly generate a top resource list, and define resource elimination rules to regularly filter the basic resource data; wherein, the high-quality resource screening rules include a popularity rating threshold, teaching feedback, and cumulative number of collections, and the resource elimination rules include a resource usage rate threshold and teaching feedback.

9. The intelligent aesthetic education curriculum resource management system according to claim 8, characterized in that: The expression of the resource heat is: U1=ω1·a+ω2·b+ω3·c+ω4·d Among them, U1 is the popularity score, ω1, ω2, ω3, and ω4 are all behavioral weights, a is the number of views, b is the number of collections, c is the number of downloads, and d is the evaluation weight; The expression of the data indicator is: Among them, U2 is the resource utilization rate, c is the total number of resource accesses, c0 is the number of invalid resource accesses, and t is the resource availability time.

10. An intelligent aesthetic education curriculum resource management system according to claim 9, characterized in that: The permission and sharing module includes: A multi-role management unit is used to define system roles, role permissions, and multi-level resource reading permissions; wherein the multi-level reading permissions include public resource reading permissions that allow all users to access and download, editable resource reading permissions that allow teachers to download, and paid resource reading permissions; An inter-school sharing unit is used to select inter-school resources from the basic resource database, define sharing conditions for the inter-school resources, sign a cooperation agreement based on the sharing conditions, and authorize resource sharing; The resource security unit is used to set the access key of the basic resource database and add watermarks and copyright ownership when previewing and downloading resources.

Images