Tech5 identifies five key principles for training of biometrics systems
Tech5 has published a list of five key principles the company uses when training and deploying its biometrics system across face, fingerprint, and iris applications, with the hope that other firms will follow their lead.
“Today we are outlining our principles for using biometric data and technology development, hoping for such rules to be adopted by other companies involved in the design and development of biometric algorithms,” says Tech5 Strategic Advisor Rob Haslam.
The first of them refers to the use of consent-based, “depersonalized” data for training of biometric algorithms, and builds on a separate announcement Tech5 made earlier this month that it has a way to build entirely consent-based facial image databases.
Secondly, the company mentioned regular training and benchmarking of the algorithms using the aforementioned data to ensure these technologies remain accurate and robust.
The second principle merges with the third: the development of technologies free from biases and “inclusive of people from all walks of life.”
The fourth key principle on Tech5’s list outlines the firm’s commitment to contribute to the development of biometrics standards worldwide with the goal of raising the level of expertise in the market.
To this end, Tech5 regularly contributes its expertise and market knowledge in its capacity as a member of various organizations, including the Biometrics Institute, the European Association for Biometrics, and ID2020.
Finally, the company renewed the importance of granting control of its biometrics technologies exclusively to certified partners.
“In order to make sure that Tech5 technologies are used to serve inclusion in all aspects of people’s lives […] we carefully select partners for the distribution of our technology components and intentionally invest in consent-based, high-quality data,” explains Tech5 CEO Machiel van der Harst.
The publication of Tech5’s key principles comes weeks after the company announced a new partnership with the Ethiopian government.
Article Topics
accuracy | algorithms | best practices | biometric data | biometrics | research and development | standards | TECH5 | training
Comments