Mazumdar, Rahimul I. (Indian Institute of Technology Guwahati), Dasgupta, Anirban (Indian Institute of Technology Guwahati), Bhowmick, Parijat (Indian Institute of Technology Guwahati)

PhantomNet: A Lightweight Convolutional Network for Distraction Detection in Automotive Drivers

Scheduled for presentation during the Video Session "On-Demand Video Presentations" (VP-VP), Saturday, November 22, 2025, 08:00−18:00, On-Demand Platform

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on April 2, 2026

Abstract

Driver distraction is a critical factor contributing to road accidents, underscoring the need for real-time, non-intrusive monitoring solutions. This paper presents PhantomNet, a lightweight convolutional neural network specifically designed for distraction classification using side-view images from the State Fram Distracted Driver Detection (SFDDD) dataset.The proposed PhantomNet introduces an efficient hybrid module that extracts expressive features with minimal computational overhead. The proposed model achieves a classification accuracy of 96.07%, outperforming existing lightweight architectures such as GhostNet, while significantly reducing model size and inference time. These results highlight PhantomNet's potential for real-time deployment in embedded automotive systems to proactively enhance the driver safety.