IDCheck – AI-based identity verification

AI-based identity verification

The AI-based IDCheck facilitates secure, automated entry management to both buildings and events, greatly speeding up the time and effort required for identity verification and document control. It reliably and quickly verifies the identity of a person by visually matching their live camera image with the identity documents required for entry. Based on privacy-compliant technologies developed by Fraunhofer IDMT, the name, date of birth and face are securely processed without the need to store or transmit data.

Robust and privacy-compliant verification 

Fraunhofer IDMT's robust image recognition technology extracts text-based personal and image information from identity documents without relying on stored biometric data (e.g. on the passport). Working in real-time, IDCheck reliably and quickly verifies permission to enter by matching a live camera image of the person to an official photo identity document. This could be a passport or employee ID card containing personal information like name and date of birth, which is then matched to another document or previously stored data, such as a personalized ticket or vaccination certificate that has no photo of the person. Depending on the security requirement, any document containing first and last name, date of birth and an official photo can be used for identity verification. The whole process runs securely on a small single-chip computer device, the size of a credit card, that does not need to be connected to a network. This helps to avoid privacy problems. There is no need to store or transfer sensitive data. 

Fast and secure entry management

IDCheck accelerates and simplifies access and entry management thanks to Fraunhofer IDMT’s automated and reliable identity verification technologies. Long waiting times at the entrance to events are significantly reduced. Additional personnel requirements for access controls and manual checks are eliminated.

Areas of Application

IDCheck can be used wherever the identity of a person needs to be verified against an official document, such as a vaccination certificate. It can be applied to many situations, such as controlling the access of employees and authorized persons to companies and institutional buildings, live events, or automated hotel check-ins.

How does IDCheck work?

IDCheck uses Fraunhofer IDMT's AI-based image analysis and face matching technology to verify an identity based on a live camera image of a person and corresponding documents required for entry, matching e.g. a personalized ticket with an official photo ID.

Features

  • Extraction of personal data (name, date of birth, face) from any identity document
  • Robust detection of personal data even when the document is placed at an angle or inaccurately
  • Extraction of facial data via a live camera on site 
  • Comparison and matching of identity data, the live image, and personal data from other documents such as vaccination certificates or personalized tickets
  • Confirmation of an "identity match" when facial and personal data correspond, and depending on the use case, an individual weighting score of facial and other characteristics is possible 
  • Privacy and data protection compliant: data is processed locally on the device; no network connection or storage of personal data is required.

Technical Information

  • Supported document camera image sizes: 1920x1080 (USB or CSI) (with a homogeneous, dark background)
  • Recommended document camera position: approx. 25 cm above the document, fixed focus with corresponding FOV. 
  • Supported live camera image sizes: 1280x720 (USB or CSI)
  • Recommended live camera position: mounting height approx. 1.6m, frontal view of the face (+/-15°), no back light
  • Duration of recognition process ~ 6s (depending on camera countdown timer settings)
  • Serial interface for transmission of comparison data, parameter settings and scores

This might also be interesting for you

 

Facial Analysis Technologies

AI-Based Facial Recognition and Analysis

 

MaskCognizer

AI-based face mask detection for public health and safety based on a Jetson nano 

 

InsightPersona

Combined face and speaker recognition to automatically analyze vast media archives

Research topic

Audio and Visual Content Analysis

Extracting meaningful data from audiovisual content

 

Research topic

Video analysis

Audio and visual content analysis