Protecting access to computers
To develop a prototype for connecting to a Windows session using facial and voice recognition within the environment of Swiss Post.
Founded in 2008 in Martigny, KeyLemon SA develops IT solutions involving facial and voice recognition. In particular, it sells facial recognition software enabling access to a Windows session without a password.
To augment its range, which included only facial recognition, KeyLemon wanted to integrate voice recognition into its products, thereby introducing a second biometric characteristic for personal identification.
A system of this kind, combining two types of biometric identification, would enable KeyLemon to target new clients who are more demanding when it comes to security, as well as to offer a ready-to-use system to large companies.
As a result of its keen interest in KeyLemon’s technology, Swiss Post agreed to participate financially in the development of a prototype of a product that combines facial and voice recognition. Swiss Post wanted to investigate an alternative system to “password login”. An innovation project was therefore initiated by The Ark, in collaboration with the Idiap Research Institute, KeyLemon and Swiss Post.
In order to satisfy the requirements of Swiss Post, KeyLemon proposed the development and implementation of an initial functional prototype. The project was divided into three stages:
a. Design and implementation of voice and facial recognition
b. Development of the prototype and integration into the Swiss Post Active Directory
c. Testing in the Swiss Post environment (around 100 employees)
In order to implement this prototype, KeyLemon had to integrate voice recognition into its current system. As it did not have all the necessary knowledge for this, the young company was able to call on assistance from the Idiap Research Institute for the development of the voice recognition aspect.
The project was completed at the end of 2012. Although it achieved its objectives, Swiss Post nevertheless does not consider the technology to be sufficiently mature. The prototype demonstrated that the technology and the algorithms should be further adapted and improved; they are not yet suitable for application on a wider scale within the company.
A second version is currently being studied. In any event the project has enabled KeyLemon to get to grips with voice identification technology, which has been introduced to its range of products. This SME will now be able to reach new markets and continue to develop.