Can LLMs Reason About Attention? Towards Zero-Shot Analysis of Multimodal Classroom Behavior

TL;DR

This paper introduces a privacy-preserving, GPU-efficient system that uses LLMs and multimodal data to analyze student attention in classrooms without storing identifiable footage.

Contribution

It presents a novel pipeline combining pose and gaze estimation with zero-shot LLM reasoning for classroom behavior analysis while ensuring privacy.

Findings

LLMs show potential in multimodal classroom behavior understanding.

The system effectively extracts and analyzes attention data without storing video footage.

Limitations exist in LLM spatial reasoning about classroom layouts.

Abstract

Understanding student engagement usually requires time-consuming manual observation or invasive recording that raises privacy concerns. We present a privacy-preserving pipeline that analyzes classroom videos to extract insights about student attention, without storing any identifiable footage. Our system runs on a single GPU, using OpenPose for skeletal extraction and Gaze-LLE for visual attention estimation. Original video frames are deleted immediately after pose extraction, thus only geometric coordinates (stored as JSON) are retained, ensuring compliance with FERPA. The extracted pose and gaze data is processed by QwQ-32B-Reasoning, which performs zero-shot analysis of student behavior across lecture segments. Instructors access results through a web dashboard featuring attention heatmaps and behavioral summaries. Our preliminary findings suggest that LLMs may show promise for multimodal behavior understanding, although they still struggle with spatial reasoning about classroom layouts. We discuss these limitations and outline directions for improving LLM spatial comprehension in educational analytics contexts.