A Multimodal Model for College English Teaching Using Text and Image Feature Extraction

Area of Science:

• Educational Technology
• Applied Linguistics
• Computer-Assisted Language Learning

Background:

• The rise of internet and multimedia technology is transforming foreign language education.
• Multimodal education is increasingly vital in English language teaching.
• Teachers are adopting multimodal teaching theories in classrooms.

Purpose of the Study:

• To investigate a multimodal model integrating text and image features for English education.
• To develop a model based on multimodal discourse theory, systemic functional linguistics, and foreign language teaching theory.
• To enhance image-text relationship evaluation through a novel multimodal approach.

Main Methods:

• Developed a multimodal model incorporating text and image features.
• Utilized multimodal discourse theory, systemic functional linguistics, and foreign language teaching theory.
• Integrated an image feature bias term into a log-bilinear natural language model.

Main Results:

• The proposed model demonstrates superior search accuracy compared to existing methods.
• The model functions as an effective image-text relationship evaluation index system.
• A trade-off was observed, with slower search speed but enhanced accuracy.

Conclusions:

• The multimodal model offers a promising advancement for English language education.
• Integrating visual and textual data improves the evaluation of multimodal content.
• Further research may focus on optimizing search speed while maintaining accuracy.