Smart Engines claims new facial recognition system uses no biometric information

Facial recognition algorithms traditionally depend on biometric data to make matches. However, with the implementation of the Artificial Intelligence Act within the European Union, the legal status of facial recognition systems has changed, especially if they pose a risk to personal freedom and privacy.

Smart Engines has identified this market gap as a need for facial recognition that does not depend on the use of biometrics. The company claims to have addressed this issue by utilizing a neural network, trained to compare facial images in a manner similar to human visual recognition.

According to Smart Engines, the AI-powered process involves three steps: image acquisition, document recognition, and face matching. The process being with acquiring a live image through a standard camera, and extracting the facial photo from an ID document. AI algorithms then compare the live image and the document image “end-to-end” to determine their similarity in a way the company compares to human matching “by eye.”

The Smart Engines highlights that its technology solves the key issue of user privacy (as defined in the AI Act) by avoiding the use of sensitive biometric information. The company believes this approach will enhance user trust, regulatory compliance and security, facilitating the widespread adoption of its technology across the European Union.

Furthermore, the demand for surveillance technologies capable of operating autonomously, without an internet connection, is growing. Smart Engines’ AI-powered solution is designed to function in such autonomous settings, promoting its wider implementation and integration across various industries.

The Smart ID Engine’s face verification software module leverages data synthesis to eliminate the bias in the training of its algorithms. This involves creating artificial data points algorithmically to support the training of machine learning modules, which is used to generate a diverse set of facial images representing a wide range of demographic groups.

The Smart ID Engine is compatible with a wide array of devices and platforms, including those based on x86 and Arm architecture, making the solution more accessible to a wide user base. It supports integration on devices running Windows, Linux, Android and iOS operating systems.

Adopting a privacy-focused approach, Smart Engines ensures that no personal and biometric information is transmitted to third-party services, nor is it saved or stored. In an effort to enhance user trust, the solution complies with leading privacy regulations such as HIPAA, GDPR, and CCPA.

The facial recognition solution uses a sustainable approach, minimizing the carbon footprint associated with AI training and deployment. The technology optimizes algorithms for real-time face verification, and the software maintains high efficiency without compromising accuracy, according to Smart Engines.

Article Topics

biometrics  |  facial verification  |  neural networks  |  research and development  |  Smart Engines