Research on Clothing Art Education Based on AI Innovation

Abstract

This study analyzes the potential of AI in fashion art education. It first addresses current challenges in education, such as monotonous teaching content, outdated methods, and disconnected teacher skills. It then develops an AI-powered application framework, combining the deep learning models ResNet152 and EfficientNetB7 for data augmentation, model optimization, and garment prediction output. This framework supports innovative teaching content design (personalized course customization and GAN-assisted creation), improved teaching methods (reinforcement learning interaction and AR virtual fitting), the development of a teaching platform (a 3D modeling virtual platform and an attention mechanism evaluation system), and the adaptation of teacher roles (AI skills training and customized loss function guidance). Experimental validation was conducted using the Prompt2Fashion dataset, employing a control group design. The results were conducted on 50 fashion art students. The results show that AI significantly enhanced students' design thinking and creativity, with an average improvement of over 20% in the experimental group, indicating high acceptance and positive feedback. However, the technical threshold needs to be optimized.