Ramco Systems builds voice biometric authentication on Azure Cognitive Services
The novel solution aims to facilitate authentication in a post-COVID world, where many countries mandate the use of facemasks, which can make facial recognition less accurate.
In addition, as the biometric tool was developed for Ramco Systems’ more than 500 large enterprises and multinational corporations, it had to satisfy specific requirements. These include user validation under two seconds, support of over 50 user profiles with registration process under 15 seconds, and enhanced privacy features.
From a technical standpoint, the solution is based upon the Azure Cognitive Services Speech to Text (STT) service, and performs ‘speaker verification’ instead of ‘speaker identification’ based on an employee saying their staff ID code.
In other words, the biometric tool has 140 voice features that it utilizes to match registered profiles. It supports multiple voice diction and tones, including those not used during registration.
The system utilizes a thermal sensor to detect human presence. The user then dictates their employee ID in front of a microphone, and the audio sample is sent to STT to ‘fetch’ the employee ID.
The app then records the user’s audio and buffers it, loads the user profile, then sends an audio sample to the “Speech Recognition” service for verification.
“With the rising concerns about data privacy in facial recognition-based solutions and the need to devise a solution for markets like the Middle East where the traditional facial recognition solutions fall short in user identification, Azure Voice biometric solution has proven to be a secure, robust alternative,” comments Ramco’s VP of Tech & Innovation Ramesh Sivasubramanian in a Microsoft blog post.
Moving forward, Ramco intends to deploy the voice biometrics platform in additional scenarios, including voice-based navigation in smartphones using the Ramco Unify mobile app, multi-user device authentication, and improved employee experience via voice-based attendance, leave requests, and other features.
“Since the solution is scalable, we are focusing on integrating it with other product features and enhancing the employee experience,” Sivasubramanian adds. “I am sure this solution can completely transform the way we interact with our enterprise systems.”