MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202631067376 A) filed by Narula Institute Of Technology on May 29, 2026, for “a Real-Time Driver’s Drowsiness Detection System Using Svm-Cnn Ensemble Model”.
Inventors include Dr. Pushpita Roy; Dr. Shubhendu Banerjee; Ankita Barua; Garima Bishnu; Priyanka Ghosh; Madhumanti Mallick; and Meghna Sengupta.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: Abstract An embedded, non-intrusive driver drowsiness detection and alert system is disclosed for enhancing automotive safety through real-time monitoring of driver fatigue. The system comprises a dual-camera hardware architecture positioned inside a vehicle cabin to capture front-facing facial expressions and upper-body postural shifts. An edge-processing unit receives the concurrent video streams and executes digital image preprocessing, facial landmark mapping, and feature isolation. A hybrid machine learning ensemble architecture processes the isolated features, combining a deep Convolutional Neural Network (CNN) for ocular feature extraction and an optimal-margin Support Vector Machine (SVM) for yawning profile classification. The predictive outputs are combined via a weighted decision fusion algorithm to calculate an accurate driver alertness index. Upon detecting structural indicators of drowsiness or microsleep, the system activates a localized alert module comprising an audible buzzer and a visual display to restore operator awareness.
Disclaimer: Curated by HT Syndication.
