Baorong Zhan, Lin Yuan, Feng Wang

Baorong Zhan, Lin Yuan, Feng Wang

Oct. 30, 2026

Digital intelligence governance, Higher education, Classroom teaching quality, Intelligent evaluation, Multi-modal data fusion.

University classroom evaluation is often assembled from questionnaires, periodic inspections, and peer observation. These sources describe overall perceptions but leave much of the within-classroom process unobserved. We formulate a multimodal evaluation scheme from the standpoint of digital-intelligence governance. The scheme separates acquisition, modality-specific processing, evidence integration, and feedback. Camera streams, classroom audio, LMS events, and environmental records are not converted directly into quality scores; they are used to derive observable descriptors such as participation events, speaking turns, response latency, instructional pacing, and room conditions. Evidence is then interpreted together with course context and human review. The design also specifies data-retention rules, role-based access, uncertainty reporting, and a route for contesting consequential decisions. Particular attention is given to temporal misalignment, missing modalities, construct validity, and discipline-dependent meaning of behavioral indicators. The resulting architecture is intended as a reference design rather than a validated scoring instrument. Its measurement properties and transferability must be examined through course-level and cross-institutional studies.

© 2026, the Authors. Published by ISCCAC

This is an open access article distributed under the CC BY-NC license