arXiv:2512.11446v1 Announce Type: new 
Abstract: Driver fatigue remains a leading cause of road accidents, with 24\% of crashes involving drowsy drivers. While yawning serves as an early behavioral indicator of fatigue, existing machine learning approaches face significant challenges due to video-annotated datasets that introduce systematic noise from coarse temporal annotations. We develop a semi-automated labeling pipeline with human-in-the-loop verification, which we apply to YawDD, enabling more accurate model training. Training the established MNasNet classifier and YOLOv11 detector architectures on YawDD+ improves frame accuracy by up to 6\% and mAP by 5\% over video-level supervision, achieving 99.34\% classification accuracy and 95.69\% detection mAP. The resulting approach deliver up to 59.8 FPS on edge AI hardware (NVIDIA Jetson Nano), confirming that enhanced data quality alone supports on-device yawning monitoring without server-side computation.

تقدم دراسة جديدة YawDD+، وهو خط أنابيب للتسمية شبه الآلي مصمم لتحسين دقة نماذج توقع التثاؤب من خلال معالجة التحديات التي تطرحها مجموعات البيانات المعلّقة بالفيديو. يحسن هذا النهج دقة الإطار بنسبة تصل إلى 6% ومتوسط الدقة (mAP) بنسبة 5%، محققًا معدلات عالية من التصنيف والكشف على أجهزة الذكاء الاصطناعي الطرفية مثل NVIDIA Jetson Nano.

Un nuevo estudio presenta YawDD+, un pipeline de etiquetado semi-automatizado diseñado para mejorar la precisión de los modelos de predicción de bostezos al abordar los desafíos que plantean los conjuntos de datos de video anotados ruidosos. El enfoque mejora la precisión de los fotogramas hasta en un 6% y la media de precisión (mAP) en un 5%, logrando altas tasas de clasificación y detección en hardware de IA en el borde como el NVIDIA Jetson Nano.

Une nouvelle étude présente YawDD+, un pipeline d'étiquetage semi-automatisé conçu pour améliorer la précision des modèles de prédiction de bâillements en s'attaquant aux défis posés par les ensembles de données vidéo annotées bruyantes. L'approche améliore la précision des images jusqu'à 6 % et la moyenne de précision (mAP) de 5 %, atteignant de hauts taux de classification et de détection sur du matériel AI en périphérie comme le NVIDIA Jetson Nano.

A new study introduces YawDD+, a semi-automated labeling pipeline designed to enhance the accuracy of yawn prediction models by addressing the challenges posed by noisy video-annotated datasets. The approach improves frame accuracy by up to 6% and mean Average Precision (mAP) by 5%, achieving high classification accuracy and detection rates on edge AI hardware like the NVIDIA Jetson Nano.

YawDD+: Frame-level Annotations for Accurate Yawn Prediction

arXiv:2601.07845v1 Announce Type: new 
Abstract: Rapid motorization in emerging economies such as India has created severe enforcement asymmetries, with over 11 million recorded violations in 2023 against a human policing density of roughly one officer per 4000 vehicles. Traditional surveillance and manual ticketing cannot scale to this magnitude, motivating the need for an autonomous, cooperative, and energy efficient edge AI perception infrastructure. This paper presents a real time roadside perception node for multi class traffic violation analytics and safety event dissemination within a connected and intelligent vehicle ecosystem. The node integrates YOLOv8 Nano for high accuracy multi object detection, DeepSORT for temporally consistent vehicle tracking, and a rule guided OCR post processing engine capable of recognizing degraded or multilingual license plates compliant with MoRTH AIS 159 and ISO 7591 visual contrast standards. Deployed on an NVIDIA Jetson Nano with a 128 core Maxwell GPU and optimized via TensorRT FP16 quantization, the system sustains 28 to 30 frames per second inference at 9.6 W, achieving 97.7 percent violation detection accuracy and 84.9 percent OCR precision across five violation classes, namely signal jumping, zebra crossing breach, wrong way driving, illegal U turn, and speeding, without manual region of interest calibration. Comparative benchmarking against YOLOv4 Tiny, PP YOLOE S, and Nano DetPlus demonstrates a 10.7 percent mean average precision gain and a 1.4 times accuracy per watt improvement. Beyond enforcement, the node publishes standardized safety events of CAM and DENM type to connected vehicles and intelligent transportation system backends via V2X protocols, demonstrating that roadside edge AI analytics can augment cooperative perception and proactive road safety management within the IEEE Intelligent Vehicles ecosystem.

تقدم دراسة جديدة عقدة إدراك Edge-AI مصممة لتحليل انتهاكات المرور في الوقت الحقيقي ونشر أحداث السلامة في الهند، مما يعالج التحديات التي تطرحها زيادة عدد المركبات وفجوة كبيرة في تطبيق القانون، مع تسجيل أكثر من 11 مليون انتهاك في عام 2023.

Un nuevo estudio presenta un nodo de percepción Edge-AI diseñado para el análisis en tiempo real de violaciones de tráfico y la difusión de eventos de seguridad en India, abordando los desafíos planteados por la rápida motorización y una significativa brecha en la aplicación de la ley, con más de 11 millones de violaciones registradas en 2023.

Une nouvelle étude présente un nœud de perception Edge-AI conçu pour l'analyse en temps réel des violations de la circulation et la diffusion d'événements de sécurité en Inde, répondant aux défis posés par la motorisation rapide et un écart d'application significatif, avec plus de 11 millions de violations enregistrées en 2023.

A new study presents an Edge-AI perception node designed for real-time traffic violation analytics and safety event dissemination in India, addressing the challenges posed by rapid motorization and a significant enforcement gap, with over 11 million violations recorded in 2023.

Edge-AI Perception Node for Cooperative Road-Safety Enforcement and Connected-Vehicle Integration

One More Thing in AI – Your Shortcut to AI Mastery

YawDD+: Frame-level Annotations for Accurate Yawn Prediction

Was this article worth reading? Share it

One More Thing in AI

LucidQuery AI

LangWatch

Dyad

Pandawhispers

Naptime

Ready to build your own newsroom?