Article: Semantic-enhanced multimodal feature fusion for English rhythm analysis Journal: International Journal of Information and Communication Technology (IJICT) 2026 Vol.27 No.81 pp.90 - 109 Abstract: Rhythm analysis of English texts is a core task in fields such as speech synthesis and affective computing. However, traditional methods often focus on single-modal data, making it difficult to capture deep semantic relationships across modalities. To address this, this paper first employs deep learning algorithms to extract multimodal English features containing significant semantic information. Subsequently, a hierarchical semantic enhancement module is designed to fuse relevant semantic information from visual and audio modalities at different text levels. Finally, we propose a multimodal feature fusion module based on cross-modal attention and gating mechanisms to suppress relatively unimportant features, thereby reducing redundant information and improving the accuracy of English text rhythm classification. Experimental results demonstrate that the proposed method achieves a weighted average F1 score of 93.17%, significantly outperforming comparison methods and enabling precise English text rhythm analysis. Inderscience Publishers - linking academia, business and industry through research

Title: Semantic-enhanced multimodal feature fusion for English rhythm analysis

Authors: Ruiqiong Yan; Shengnan Sun

Addresses: School of Foreign Languages, Zhongyuan Institute of Science and Technology, Xuchang, 461000, China ' School of Foreign Languages, Zhongyuan Institute of Science and Technology, Xuchang, 461000, China

Abstract: Rhythm analysis of English texts is a core task in fields such as speech synthesis and affective computing. However, traditional methods often focus on single-modal data, making it difficult to capture deep semantic relationships across modalities. To address this, this paper first employs deep learning algorithms to extract multimodal English features containing significant semantic information. Subsequently, a hierarchical semantic enhancement module is designed to fuse relevant semantic information from visual and audio modalities at different text levels. Finally, we propose a multimodal feature fusion module based on cross-modal attention and gating mechanisms to suppress relatively unimportant features, thereby reducing redundant information and improving the accuracy of English text rhythm classification. Experimental results demonstrate that the proposed method achieves a weighted average F1 score of 93.17%, significantly outperforming comparison methods and enabling precise English text rhythm analysis.

Keywords: English text; rhythm classification; semantic enhancement; multimodal fusion; deep learning.

DOI: 10.1504/IJICT.2026.155111

International Journal of Information and Communication Technology, 2026 Vol.27 No.81, pp.90 - 109

Received: 24 Mar 2026
Accepted: 01 May 2026
Published online: 27 Jul 2026 *

Title: Semantic-enhanced multimodal feature fusion for English rhythm analysis

Keep up-to-date