arXiv:2206.13356v2 Announce Type: replace 
Abstract: Online exams, conducted via video conferencing platforms such as Zoom, have become popular in educational institutions since COVID-19. While convenient, ensuring the integrity and security of online exams remains challenging, as traditional invigilation struggles to effectively monitor multiple student video feeds in real time. In this paper, we present iExam, an effective online exam proctoring and analysis system that combines lightweight face detection and deep recognition. iExam employs real-time face detection to assist invigilators in continuously monitoring student presence, and leverages deep face recognition for post-exam video analysis to identify abnormal behaviors--including face disappearance, face rotation, and identity substitution. To realize this system, we address three core challenges: (i) designing a lightweight approach to efficiently capture and analyze exam video streams in real time; (ii) developing an enhanced OCR method to automatically extract student identities from dynamically positioned Zoom name tags, enabling reliable ground truth labeling without manual intervention; and (iii) optimizing the training and inference pipeline to significantlyreduce resource and time requirements on ordinary teacher devices. Extensive experiments demonstrate that iExam achieves 90.4% accuracy for real-time face detection and 98.4% accuracy for post-exam face recognition, while maintaining low overhead. These results show that iExam can substantially enhance the automation and reliability of online exam proctoring in practice.

تم تطوير نظام جديد لمراقبة الامتحانات عبر الإنترنت يسمى iExam، والذي يدمج تقنيات الكشف عن الوجه الخفيفة والتعرف العميق لتعزيز نزاهة التقييمات عبر الإنترنت التي تتم عبر منصات مثل Zoom. يتناول هذا النظام تحديات مراقبة عدة تدفقات فيديو للطلاب في الوقت الفعلي ويتضمن ميزات للتحليل بعد الامتحان للكشف عن سلوكيات غير طبيعية مثل اختفاء الوجه واستبدال الهوية.

Se ha desarrollado un nuevo sistema de supervisión de exámenes en línea llamado iExam, que integra tecnologías de detección facial ligera y reconocimiento profundo para mejorar la integridad de las evaluaciones en línea realizadas a través de plataformas como Zoom. Este sistema aborda los desafíos de monitorear múltiples transmisiones de video de estudiantes en tiempo real e incluye características para el análisis posterior al examen para detectar comportamientos anormales como la desaparición de rostros y la sustitución de identidad.

Un nouveau système de surveillance des examens en ligne, nommé iExam, a été développé, intégrant des technologies de détection faciale légère et de reconnaissance approfondie pour améliorer l'intégrité des évaluations en ligne réalisées via des plateformes comme Zoom. Ce système répond aux défis de la surveillance de plusieurs flux vidéo d'étudiants en temps réel et comprend des fonctionnalités d'analyse post-examen pour détecter des comportements anormaux tels que la disparition du visage et la substitution d'identité.

A new online exam proctoring system named iExam has been developed, which integrates lightweight face detection and deep recognition technologies to enhance the integrity of online assessments conducted via platforms like Zoom. This system addresses the challenges of monitoring multiple student video feeds in real time and includes features for post-exam analysis to detect abnormal behaviors such as face disappearance and identity substitution.

Effective Online Exam Proctoring by Combining Lightweight Face Detection and Deep Recognition

<A HREF="https://www.zoom.com/en/blog/humanitys-last-exam-zoom-ai-breakthrough/"><IMG VSPACE="4" HSPACE="4" BORDER="0" ALIGN="RIGHT" SRC="http://www.techmeme.com/251212/i27.jpg"></A>
<A HREF="http://www.techmeme.com/251212/p27#a251212p27" TITLE="Techmeme permalink"><IMG WIDTH=11 HEIGHT=12 SRC="http://www.techmeme.com/img/pml.png" STYLE="border:none;padding:0;margin:0;"></A> Xuedong Huang / <A HREF="https://www.zoom.com/">Zoom</A>: 
<A HREF="https://www.zoom.com/en/blog/humanitys-last-exam-zoom-ai-breakthrough/">Zoom says its &ldquo;federated AI&rdquo; model, combining its SLM with open- and closed-source models, got 48.1% on Humanity's Last Exam vs. 45.8% for Gemini 3 Pro w/ tools</A>&nbsp; &mdash;&nbsp; Federated innovation driving breakthrough results in complex AI testing&nbsp; &mdash;&nbsp; In this blog&nbsp; &mdash;&nbsp; 01 Understanding the HLE challenge

أعلنت Zoom أن نموذج الذكاء الاصطناعي الفيدرالي الخاص بها حقق درجة 48.1% في امتحان الإنسانية الأخير، متفوقًا على نموذج Gemini 3 Pro من Google الذي حصل على 45.8%. يجمع هذا النموذج بين SLM الخاص بـ Zoom ونماذج مفتوحة ومغلقة المصدر، مما يظهر تقدمًا كبيرًا في قدرات الذكاء الاصطناعي.

Zoom anunció que su modelo de IA federada logró un puntaje del 48.1% en el Examen Final de la Humanidad, superando al Gemini 3 Pro de Google, que obtuvo un 45.8%. Este modelo combina el SLM de Zoom con tecnologías de IA de código abierto y cerrado, mostrando un avance significativo en las capacidades de IA.

Zoom a annoncé que son modèle d'IA fédéré a obtenu un score de 48,1 % à l'examen final de l'humanité, surpassant le Gemini 3 Pro de Google, qui a obtenu 45,8 %. Ce modèle intègre le SLM de Zoom avec des technologies d'IA à code ouvert et fermé, montrant une avancée significative dans les capacités de l'IA.

Zoom announced that its federated AI model achieved a score of 48.1% on Humanity's Last Exam, outperforming Google's Gemini 3 Pro, which scored 45.8%. This model integrates Zoom's SLM with both open- and closed-source AI technologies, showcasing a significant advancement in AI capabilities.

Zoom says its "federated AI" model, combining its SLM with open- and closed-source models, got 48.1% on Humanity's Last Exam vs. 45.8% for Gemini 3 Pro w/ tools (Xuedong Huang/Zoom)

Effective Online Exam Proctoring by Combining Lightweight Face Detection and Deep Recognition

Was this article worth reading? Share it

AiReelGenerator.com

Pizi AI

BlitzToksAI

ComfyUI

OpExams

TheQuizMaster

Ready to build your own newsroom?