Facial Recognition Solutions – Ready for the Future

Vara has collaborated with STC Innovations, a leading developer of facial recognition, voice and multimodal biometric systems to come up with revolutionary security analysis solutions. We deploy high definition video records and state-of-the-art facial recognition technology to allow Security, Law Enforcement and Investigation Agencies to analyze the person and his document live, without coercion, and in great detail.

• The aviation industry serves more than 4.5 billion passengers every year, making it an easy target for attacks by international/domestic or extremist groups.
• An airport easily consists of 1000s of cameras, generating a huge volume of data. Manning multiple control rooms and achieving effective oversight requires significant manpower and is too expensive to sustain in the long term.
• Prior to the pandemic, the Mumbai Metro had a daily ridership of 4,50,000 people while the Delhi Metro carried 50,64,000 people every day. The Mumbai Metro has deployed a total of 700 CCTV cameras and the Delhi Metro has around 5,200 CCTV cameras across stations.
• Without the right tools, manual inspection of all these cameras is practically impossible.

We’ve partnered with the industry leader, STC Innovations to develop a product application that uses the available video surveillance systems i.e. existing cameras as tools providing us with a leg up using the existing infrastructure. The cameras in the surveillance system provide raw data feed with the faces of passengers, storing information, identifying and processing images in a manner that is similar to the human eye. This data is analysed using DSP (Digital Signal Processing) and Computer Vision with Deep Learning (part of the Machine learning paradigm) to trace facial patterns and identify people, as well as suspicious activities.

SMART TRACKER

Rapid Identification For Security and Queue Management

To understand the system practically, let us look at how our implementation works.

• Face Tracking
  The video stream is shared in the system and faces are recognized and tagged. These are then identified and compared using specialised biometrics with faces in the current database.
• Personal Identification
  Each person is identified with their name and unique identity.
• Storage and Analysis
  After identification, the system alerts and tracks individuals at other events and checks for malicious activity and blacklists any leads.
• Advanced Recognition
  The system employs a multipronged approach with recognition extended to partially covered faces. With the advent of covid 19 and people wearing masks, we found it to be a key feature for building a future-ready surveillance application.

Eyeballs captured

Secondly, the ability to differentiate live faces from fake ones. This comes to play when any suspicious individual tries to cheat the system.

Facial autonomy captured

Thirdly, overcoming bad lighting conditions to present better results. All the corners of an airport might not be well lit and therefore, this feature provides better results in case of any rising eventualities.

CCTV footage of a passenger in dull light capturing the silhouette

• Notification Alerts
  If there are any questionable identities found in the database, they are tagged and their faces are forwarded to the airport security staff and customs.
• Reduced Reaction Time
  The application of this system reduces reaction time manifold thereby reducing the pressure on the ground airport staff and easing tensions of frequent and first-time flyers.
• Hands-free Passengers
  With the failsafe biometrics system, customers do not need to carry their passes in their hands. Identification takes place exclusively via technical means via three possible ways –
  • Combination of passenger profile with a biometric profile of a person enables online check-in by adding a photo when registering via a mobile or web application.
  • Taking pictures for instant authorisation at the counter reception to get you through.
  • A different use case is at the terminal where photography enables automatic registration. For Domestic Flights we can enable identity confirmation using face image.

• Database (up to several millions cards) containing face images and metadata
• AI technology and identity-verification algorithm allow extraction of real-time biometric pattern of the person
• Assisting security staff and decreasing human error in current processes
• Integration with other urban facilities like transport, municipal services, e-government etc.

• 100% faces search success rate
• IA algorithm precision: 99.99%
• Identification time: Under 3 seconds
• Biometric scoring precision: 94.76 %
• Search through 1,000,000 samples database in 3 seconds
• Tracking 15+ faces in one frame

Who Are We: Vara Technology is the front runner in embracing innovation and leveraging world class technology to deliver maximum business value to organizations. Headquartered in Kolkata, Vara is dedicated towards bringing change through technology with long-term and sustainable objectives. We develop smart solutions that drive new value to help organizations optimize their service delivery.