Article: Classroom student emotion recognition using an improved segmentation clustering and multi-feature fusion emotion recognition algorithm Journal: International Journal of Continuing Engineering Education and Life-Long Learning (IJCEELL) 2026 Vol.36 No.10 pp.1 - 26 Abstract: To address noise, speech masking, and weak robustness in classroom emotion recognition, this study proposes a model combining enhanced segmentation clustering and multi-feature fusion. An improved U-Net with local loss supervision first performs denoising. Secondly, using MFCC features combined with Bayesian segmentation and K-means clustering to process speech signals. Finally, MFCC, formant, and pitch features are integrated into an attention-based BiLSTM for emotion recognition. Results show the U-Net achieved a loss of 0.26 after 16 iterations, with PESQ at 3.01 and STOI 84.21%. Segmentation false negative and positive rates were 13.84% and 12.52%. K-means purity reached 90.85%. The multi-feature model attained 92.98% accuracy for excited emotion, and the full system reached 93.62% test accuracy. The model improves recognition in complex classrooms, supporting personalised smart education. Inderscience Publishers - linking academia, business and industry through research

Title: Classroom student emotion recognition using an improved segmentation clustering and multi-feature fusion emotion recognition algorithm

Authors: Xiaohong Wang

Addresses: School of Humanities, Weinan Normal University, Weinan, 714099, China

Abstract: To address noise, speech masking, and weak robustness in classroom emotion recognition, this study proposes a model combining enhanced segmentation clustering and multi-feature fusion. An improved U-Net with local loss supervision first performs denoising. Secondly, using MFCC features combined with Bayesian segmentation and K-means clustering to process speech signals. Finally, MFCC, formant, and pitch features are integrated into an attention-based BiLSTM for emotion recognition. Results show the U-Net achieved a loss of 0.26 after 16 iterations, with PESQ at 3.01 and STOI 84.21%. Segmentation false negative and positive rates were 13.84% and 12.52%. K-means purity reached 90.85%. The multi-feature model attained 92.98% accuracy for excited emotion, and the full system reached 93.62% test accuracy. The model improves recognition in complex classrooms, supporting personalised smart education.

Keywords: emotion recognition; speech signal processing; segmentation and clustering; multi-feature fusion; MFF; support vector machines; SVMs; classroom engagement; deep learning; attention mechanism.

DOI: 10.1504/IJCEELL.2026.154217

International Journal of Continuing Engineering Education and Life-Long Learning, 2026 Vol.36 No.10, pp.1 - 26

Received: 15 Sep 2025
Accepted: 04 Feb 2026
Published online: 16 Jun 2026 *

Title: Classroom student emotion recognition using an improved segmentation clustering and multi-feature fusion emotion recognition algorithm

Keep up-to-date