Related Experiment Video
Updated: Aug 6, 2026
06:37
Artificial Intelligence-Based System for Detecting Attention Levels in Students
Published on: December 15, 2023
StudentNet: an edge-deployable approach for student behavior detection in smart classrooms
Frontiers in Computational Neuroscience
|July 24, 2026
Summary
StudentNet enhances student behavior detection in classrooms, improving accuracy in complex scenes. This novel framework is suitable for real-time smart classroom applications on edge devices.
Area of Science:
- Computer Vision
- Artificial Intelligence
- Educational Technology
Background:
- Accurate student behavior identification is crucial for intelligent education systems.
- Existing methods struggle with complex backgrounds and crowded scenes, leading to reduced detection accuracy.
Purpose of the Study:
- To develop a novel framework, StudentNet, to improve student behavior detection in challenging classroom environments.
- To address limitations of existing methods in handling complex backgrounds and high-density scenarios.
Main Methods:
- StudentNet incorporates the Behavior-Enhanced Fusion Super-Resolution Network (BEF-SRNet) for feature discriminability.
- The Direction-Sensitive Adaptive Convolution (DSAC) module extracts structural cues and suppresses noise.
- Adaptive Spatial Feature Fusion (ASFF) strategy enhances scale-invariant feature representation.
Main Results:
- StudentNet improved mAP@50 by 6.3% and F1-score by 6.4% compared to YOLOv8n on the SCB-Dataset3.
- The INT8-quantized model achieved 38.4 FPS on the Core-3399Pro-JD4 edge platform with minimal accuracy loss.
Conclusions:
- StudentNet significantly enhances student behavior detection accuracy in complex classroom settings.
- The framework demonstrates suitability for real-time smart classroom applications on edge devices.