Integrating Biotechnology Approaches to Optimize Algorithms for Detecting Students' Classroom Behavior

Yao Fu
College of Medicine, Jingchu University of Technology, Jingmen, 448000, China
Zhang Lei
College of Medicine, Jingchu University of Technology, Jingmen, 448000, China

DOI:https://doi.org/10.5912/jcb2517

Abstract:

Advancements in biotechnology are increasingly influencing data analytics across sectors, including educational technology. This study introduces a biotechnology-driven optimization of classroom behavior detection algorithms, utilizing bio-inspired neural architectures to enhance the analysis of student engagement. The research adapts the Single Shot MultiBox Detector (SSD) framework with biologically motivated design principles, aiming to improve real-time biometric behavior recognition. The optimized SSD algorithm integrates reverse residual structures inspired by neural plasticity to minimize information loss during feature extraction. Additionally, high-resolution feature fusion mechanisms—mimicking hierarchical visual processing in the human cortex—enable the accurate detection of subtle behavioral cues such as posture shifts and micro-expressions, which are key indicators of engagement. Experimental results demonstrate a 12.12% increase in recognition accuracy (90.21% compared to 80.46% for CNN baselines), validating the potential of bio-inspired network architectures in capturing nuanced behavioral patterns. This research contributes to the biotechnology field by proposing a scalable framework that integrates AI and biometric analysis, offering an innovative approach for enhancing educational assessments while adhering to privacy standards.