Exécution de notre outil de synthèse sur un nouvel article

Vérifiez l’article ci-dessous, puis lancez sa synthèse.


HARM-YOLO11n: A Lightweight Real-Time Object Detection Framework for Robust Classroom Behavior Monitoring

Zihang Zhang (Xi'an Polytechnic University), Yonghua Mao (Xi'an Polytechnic University), Yingcang Ma (Xi'an Polytechnic University), Xiaolin Gui (Xi'an Jiaotong University)

Sensors
DOI: 10.3390/s26154770 (Computer Science)
Licence : CC BY

Résumé : Real-time classroom behavior monitoring is an important component of intelligent education systems, yet it remains challenging due to dense student layouts, frequent occlusions, and small-scale behavior targets. To address these challenges, this paper proposes Hierarchical Adaptive Re-parameterized Multi-scale YOLO11n (HARM-YOLO11n), a lightweight object detection framework based on YOLO11n for classroom behavior analysis. The proposed framework integrates structural re-parameterized feature enhancement (SRFE), a spatial-channel adaptive module (SCAM), position-adaptive multi-scale fusion (PAMSF), and IoU-adaptive soft weight (IASW) to improve feature representation, multi-scale interaction, and optimization behavior. SRFE enhances feature extraction through structural re-parameterization, the SCAM reduces redundant feature responses via selective spatial-channel computation, PAMSF performs adaptive multi-scale feature aggregation, and IASW introduces localization-aware sample weighting during training. Experiments on POCO-Dataset, SCB-Dataset3, and STBD-08 show that HARM-YOLO11n achieves improved detection performance compared with representative lightweight detectors while maintaining real-time inference efficiency. Multi-dataset evaluations further indicate stable performance under different data distributions. The results suggest that the proposed framework provides a practical approach for classroom behavior detection in resource-constrained educational environments.

Publié le 27 Jul. 2026