Multimodal deep learning for art behavior analysis and personalized teaching path generation

Abstract In the current era of continuous innovation in educational philosophy, the importance of art education for students ‘creativity, aesthetic ability, and all-round development has become increasingly prominent, with personalized education becoming a key trend. This study focuses on art educat...

Full description

Saved in:
Bibliographic Details
Main Authors: Yikun Li, Jie Shi
Format: Article
Language:English
Published: Springer 2025-08-01
Series:Discover Artificial Intelligence
Subjects:
Online Access:https://doi.org/10.1007/s44163-025-00480-w
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849226077000433664
author Yikun Li
Jie Shi
author_facet Yikun Li
Jie Shi
author_sort Yikun Li
collection DOAJ
description Abstract In the current era of continuous innovation in educational philosophy, the importance of art education for students ‘creativity, aesthetic ability, and all-round development has become increasingly prominent, with personalized education becoming a key trend. This study focuses on art education, aiming to address the lack of personalization in traditional teaching methods by proposing innovative solutions using multimodal deep learning technology. The research constructs a Multimodal Feature Fusion Network (MFFN) and an Evolutionary Path Generation Algorithm (EPG), utilizing Kinect sensors, digital tablets, and other devices to collect multi-modal data including visual, tactile, behavioral, and environmental information, establishing a large-scale art learning dataset. Experimental results show that the MFFN model performs excellently in technique recognition, emotion judgment, and attention detection tasks, achieving an accuracy rate of 93.7% in technique recognition. Personalized teaching paths generated based on EPG can dynamically adjust according to learners’ real-time capability profiles and learning progress, improving learning efficiency by 42.6%. Additionally, the study constructs a system architecture covering data layers, algorithm layers, and application layers, enabling real-time dynamic adjustment and visual traceability of teaching paths. This research establishes a multimodal analysis paradigm for art behavior, promoting the digitalization and personalization of art education, with significant application value in areas such as art training institutions, special education, and the digital transformation of art education.
format Article
id doaj-art-8d5e8aaf0eca4e9a9cd5aca66bdf6bb3
institution Kabale University
issn 2731-0809
language English
publishDate 2025-08-01
publisher Springer
record_format Article
series Discover Artificial Intelligence
spelling doaj-art-8d5e8aaf0eca4e9a9cd5aca66bdf6bb32025-08-24T11:39:57ZengSpringerDiscover Artificial Intelligence2731-08092025-08-015111810.1007/s44163-025-00480-wMultimodal deep learning for art behavior analysis and personalized teaching path generationYikun Li0Jie Shi1School of Arts, Weinan Normal UniversitySchool of Arts, Weinan Normal UniversityAbstract In the current era of continuous innovation in educational philosophy, the importance of art education for students ‘creativity, aesthetic ability, and all-round development has become increasingly prominent, with personalized education becoming a key trend. This study focuses on art education, aiming to address the lack of personalization in traditional teaching methods by proposing innovative solutions using multimodal deep learning technology. The research constructs a Multimodal Feature Fusion Network (MFFN) and an Evolutionary Path Generation Algorithm (EPG), utilizing Kinect sensors, digital tablets, and other devices to collect multi-modal data including visual, tactile, behavioral, and environmental information, establishing a large-scale art learning dataset. Experimental results show that the MFFN model performs excellently in technique recognition, emotion judgment, and attention detection tasks, achieving an accuracy rate of 93.7% in technique recognition. Personalized teaching paths generated based on EPG can dynamically adjust according to learners’ real-time capability profiles and learning progress, improving learning efficiency by 42.6%. Additionally, the study constructs a system architecture covering data layers, algorithm layers, and application layers, enabling real-time dynamic adjustment and visual traceability of teaching paths. This research establishes a multimodal analysis paradigm for art behavior, promoting the digitalization and personalization of art education, with significant application value in areas such as art training institutions, special education, and the digital transformation of art education.https://doi.org/10.1007/s44163-025-00480-wMultimodal deep learningArt learning behavior analysisPersonalized teaching pathEvolutionary algorithmMultimodal data fusion
spellingShingle Yikun Li
Jie Shi
Multimodal deep learning for art behavior analysis and personalized teaching path generation
Discover Artificial Intelligence
Multimodal deep learning
Art learning behavior analysis
Personalized teaching path
Evolutionary algorithm
Multimodal data fusion
title Multimodal deep learning for art behavior analysis and personalized teaching path generation
title_full Multimodal deep learning for art behavior analysis and personalized teaching path generation
title_fullStr Multimodal deep learning for art behavior analysis and personalized teaching path generation
title_full_unstemmed Multimodal deep learning for art behavior analysis and personalized teaching path generation
title_short Multimodal deep learning for art behavior analysis and personalized teaching path generation
title_sort multimodal deep learning for art behavior analysis and personalized teaching path generation
topic Multimodal deep learning
Art learning behavior analysis
Personalized teaching path
Evolutionary algorithm
Multimodal data fusion
url https://doi.org/10.1007/s44163-025-00480-w
work_keys_str_mv AT yikunli multimodaldeeplearningforartbehavioranalysisandpersonalizedteachingpathgeneration
AT jieshi multimodaldeeplearningforartbehavioranalysisandpersonalizedteachingpathgeneration