iB3

App

Advanced and Accurate 2D & 3D Face Recognition Software Technology

When it comes to delivering the highest performance level in facial recognition, iB3 offers nothing less than the best. While meeting the international standards, there are two primary applications for our software, identity verification and access controls.

iB3 is a leading supplier for ALL your security needs.

Our low-resolution and high-resolution facial recognition software products have a direct application in a large number of business channels some of which are:

  • Personal Identification: Numerous applications in security, employment, personal banking etc. where identification can be easily implemented and data checked instantaneously.
  • Border Security Control/Law Enforcement: Facial recognition is the fastest and most accurate method to assist in identity verification.
  • Retail:Within seconds a known shoplifter can be detected saving the storeowner thousands of dollars without having to enlarge the security staff.
  • Transportation: Airports, railroad stations, cruise lines- rapid registration and identification of legitimate passengers and identify security risks instantaneously.
  • Residential/Commercial Real Estate: Identify owners and tenants and register visitors in secure and guarded properties using facial recognition.
  • Hospitals and Pharmacies: Enroll patients using facial recognition and eliminate fraud and identification errors.
  • Medical Records:BioFace-ID advanced facial recognition technology implements a higher standard security and compliance.
  • Medicare/Medicaid Verification: Eliminates identity theft and billing errors by matching the subjects’ face to their records.
  • Insurance Verification: Absolutely identify insured individuals their beneficiaries by establishing a database of their faces.
  • Military: Personnel databases using facial recognition technology can have authorization levels for increased security at various military and governmental sites.
  • Credit Card and ATM Authorization: Eliminates ID theft since the face is identified in real time rather than a PIN.
  • E-commerce ID Verification: Reduces fraud by absolutely identifying the user.
  • Department of Motor Vehicles: Augments license and vehicle registration information by quickly confirming the drivers’ identity.
  • Federal and State Applications: As a rapid identification tool for many governmental programs where subject registration is required.
  • Entertainment/Gaming: Use zone screening and crowd surveillance facial recognition to find known unwanted individuals in large crowds to rapidly alert security forces to that persons’ presence.
  • Manufacturing: Maintain secure “factory floor” environments by facial recognition screening of employees and visitors.
  • OEM: video surveillance, cameras, door, locking devices, computers, smart phones, apps

… and many, many more…

We deliver user-friendly software that offers unlimited options making it adaptable to all needs worldwide.

  • Retrofit: With a multitude of options, we have the capability of retrofitting to existing products. While it is not always feasible for all clients to fully transition into a new system, under certain conditions we are able to adapt to their existing equipment, allowing them to benefit from the new technology.
  • OEM Embedding: We understand the importance of the original equipment manufacturer staying competitive. Our company accomplishes this by using embedded software technology. We work with our clients to create a software feature that will set their product apart and above all the rest. An example of embedded software through manufacturers is a telephone built into a car.

Low Resolution Software Technology. Reliable, Accurate, Dependable.

Face recognition or face verification processes require:

  • Face detection: Information on the position and size of each face visible in the image. The detection module is extremely fast due to the proprietary boosting algorithm with weak classifiers.
  • Face normalization: process that marks points of morphological interest throughout the face area (eyes, mouth, and eyebrows) in order to properly rotate and scale the image. This operation moves all these points to a predefined position. Due to our proprietary implementation of the Statistical Models of Appearance method, our system keeps functioning correctly, even in the presence of severe occlusions (dark sunglasses, scarves, etc.).
  • Feature extraction: A selection of features are extracted from the normalized image in order to maximize the robustness against environmental disturbance and noise, such as non-optimal pose, non-neutral facial expressions and variable illumination conditions. The implementation of this module involves advanced filtering techniques of linear and non-linear nature.
  • Building of the biometric template: The extracted feature vector template on the previous step is fed to a statistical analysis engine. Using the trained statistical models, the feature vector is reduced to a smaller one that optimally describes the person’s identity.
  • Biometric template comparison: Face recognition and face verification process are based on measuring template differences. The normalized template distance is the similarity value of the compared identities.

iB3 Access

Specifically designed APP for cooperative recognition applications:

  • Access control for public and governamental areas (borders, hospitals, public offices)
  • Parking control

Dedicated enrolment stations

  • Different HW configurations

Zone screening (ZS)

Specifically designed APP for non cooperative recognition applications:

  • Stadiums
  • Public places
  • Stations
  • Airports

How to identify a face among tens of thousands of people

Face recognition is the most effective biometric technique for overcrowded zones (airports, rail stations, etc.) because it relies on distant cameras to identify people.

The identification abilities sustained by video cameras would be a daunting task for human investigators trying to remember several hundred faces and match them against tens of thousands of faces per day.

The system can be used as a screening method that selects only those people that are worth double checking: the final decision on their identity can then be made by an operator (security monitor, tablet, smartphone, etc.)

USE CASE: non-collaborative subject

Case: non-collaborative moving subject

  • Blurred image
  • Reduced spatical resolution (loss of details)

HW solutions:

  • Project parameter: distance between the center of the eyes 50-100 pixels
  • High-resolution cameras (1.5 Mp – 5 Mp)
  • Area coverage with 2 or more side by side cameras

SW solutions:

  • High-speed image acquisition
  • Quality index per face
  • Identity tracking (reconstruction of faces by using successive frames)
  • Selection of best image of each identity

iB3 Market

Specifically designed APP for marketing applications

  • Department stores
  • Airport shops
  • Fashion boutiques
  • Franchise networks

Common features

Modularity

  • Creation of specific sequence of algorithms using an encrypted XML file with full description of each step of the solution
  • Capability to modify the sequence “on-the-fly” in order to adapt to specific situations or new challenges

Personalised training with in depth representation of the image. Definition of gender, race, age and pose