Skip to content
V

Visum

AI-powered drowsiness detection DASHCAM

Created on 29th June 2025

V

Visum

AI-powered drowsiness detection DASHCAM

The problem Visum solves

Every single day, India loses over 500 lives to road accidents with, driver drowsiness and inattention accounting for nearly 40% of these fatalities, as per recent reports. The problem is even more severe among older vehicles and commercial fleets, which lack built-in Driver Monitoring Systems (DMS). Despite the rising number of accidents on highways like the Yamuna Expressway—where over 44% of crashes between 2012 and 2023 were caused by drivers dozing off alone—preventive technologies remain underutilized. our solution addresses this gap through AI-driven real-time monitoring, offering an affordable, and life-saving solution.

Our solution (Visum) is an AI-powered, plug-and-play Driver Monitoring System designed to reduce drowsiness-induced road accidents, especially in legacy vehicles lacking modern safety features. Using advanced computer vision, real-time sensor inputs, and Research-backed algorithms, it tracks critical indicators like eye aspect ratio, yawning frequency, and head posture to monitor driver alertness. If signs of fatigue are detected, Visum triggers instant alerts or activates safety protocols to prevent accidents. Designed for the Indian market, it complies with AIS-184 standards and is ideal for trucks, buses, and commercial fleets. Visum offers universal compatibility, easy installation, and is backed by proven research and impactful innovation.

Challenges we ran into

One major challenge we faced while building Visum was minimizing false alarms caused by natural head movements and facial variations among different drivers. Early versions often misclassified normal behavior—like mirror glances or slight head tilts—as drowsiness, leading to unnecessary alerts. We resolved this by training on a diverse dataset, integrating temporal analysis, and adding contextual filters to differentiate genuine fatigue from regular movements. Another key hurdle was ensuring stable cloud connectivity for our dashboard website, especially in remote highway areas with poor internet. We tackled this by implementing local data caching and automatic sync when a stable connection is available, ensuring uninterrupted monitoring and reliable data access.

Discussion

Builders also viewed

See more projects on Devfolio