Cheating Detection Pipeline for Online Interviews

Abstract

Global precautions against the pandemic made the online meeting systems widespread. Most of the companies and academic institutions utilize these systems for their recruitment processes and also for online exams. This led to the integration of anti-cheating analysis becoming a necessity for online meetings. We built an ideal pipeline for such an anti-cheating system and designed its components. These components may vary depending on use cases. However, some basic functionalities must remain for proper software. The pipeline consists of Face detection, Face recognition, Object detection, Face tracking, and Result analysis. We evaluated this pipeline and its components on a private interview video dataset. The dataset is labeled by its cheating status for overall video and also by the presence of individual cheating events. We utilized faster implementations of squeezed versions of up-to-date deep learning detection models to be able to process videos faster. Ultimately, our pipeline presents a guideline to detect and analyze cheating activities in an online interview video efficiently.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 19 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.