A Major European Railway Operator Achieved 94% Accuracy in Real-Time Train Door Detection Using AI-Powered Accessibility Solution
Overview
A major European railway operator serving millions of passengers annually, committed to delivering safe, inclusive, and innovative travel experiences. The organization is at the forefront of adopting technology to improve accessibility, operational efficiency, and passenger satisfaction.
The railway operator faced accessibility gaps for visually impaired passengers in identifying train doors. Traditional methods lacked innovation and failed to provide real-time, personalized assistance. Key issues included:
Limited independence and inclusivity for visually impaired passengers.
No automated solution for real-time detection of train door locations.
Missed opportunities to leverage AI for improving accessibility, safety, and customer experience.
Solution
Coforge developed a video analytics-driven accessibility solution using deep learning and mobile integration to empower visually impaired passengers:
Explored the feasibility of combining video analytics, deep learning, and mobile-based inference to identify train doors in real-time.
Trained neural networks on AWS GPUs using Python, Keras, and TensorFlow for high-accuracy detection.
Implemented on-device inference on iPhones (edge analytics) for instant recognition without internet dependency.
Delivered tactile feedback via mobile app to alert passengers when approaching open train doors.
The Impact
94% accuracy in recognizing open train doors of specific models in real time.
100% real-time tactile alerts to improve passenger safety and independence.
Seamless mobile app experience enabling accessibility without connectivity issues.
