nir-vana innovation network

The Turkish company is located in Ankara, Turkey. It was founded in 2019 with an aim for artificial intelligence (AI) and spreading its benefits to society. The company’s founders strongly believe in an AI-centric future that withholds unprecedented benefits to businesses that adopt and embrace advances in AI and machine learning (ML) in a timely manner. As an integral part of this vision, they dwell on interpreting AI-centric software with a different perspective and grasping how AI and ML ideas are applied in software or hardware along with their theoretical foundations.

The company’s mission is to put evolving AI and ML technologies into technical context and make them accessible to interested parties who aim to develop high technology products as well as developing highly tailored and optimized AI-centric software in a ready to use format. The offered product is a real-time face recognition and identification system aimed at security, surveillance, and access control applications, capable of serving the public security needs of national security forces as well as the facility monitoring needs of public and private institutions. The system software is capable of recognizing and identifying face pictures of an individual without strict limitations on resolution and lighting conditions in real-time. Allowed database sizes, still keeping real-time operability, can go up to hundreds of millions of persons with single or multiple face pictures for each individual. Furthermore, the system is robust against varying facial expressions, aging, changing hairstyles, presence of beard and/or mustache, glasses, headwear, and other cosmetic variations.

The developed face recognition system can be configured to run completely on the customer’s local servers without requiring any operations on cloud servers. This helps protect personal biometric data from any external attacks. To further increase the security of personal biometric data, the creation and updating of personal databases are left to the customer through an easy-to-use interface. The system is designed in such a way that the user can easily customize system use by categorizing its personnel or suspicious persons in its own private database. The offered system may be applied in access control of corporate buildings, warehouses, factories, varying forms of transportation (airplanes, trains, hotels, hospitals, and highway ID checkpoints).

The company is looking for partners to work with under the license agreement model or the commercial agency agreement model. The partners can be SMEs, large corporations, or government agencies. In the case of the licensing agreement, the interested partner(s) will be provided with the operational executable of the offered face recognition and identification system, detailed documents, and necessary information on how to deploy and operate the system for a one-time upfront payment and a license-fee for each operational instance of the system. In the case of commercial agency agreement with technical assistance, the partner interested in the offered face recognition and identification system will be provided with the operational executable of the offered system and training concerning its usage, adoption to different use cases, or additional features.

Advantages & innovations

Without strict limitations on input picture resolution, the system software is capable of recognizing and identifying face pictures of an individual. Database sizes that keep real-time operability, can go up to hundreds of millions of people with single or multiple face pictures for each individual. For example, given a single face picture of an individual, the system is capable of inquiring a database of 100 million persons with one face picture for each person under 500 milliseconds. The system is based on a novel processing pipeline that consists of multiple deep neural networks for input image pre-conditioning and face processing. As a result, it is robust against cosmetic variations and lighting conditions such as:  Varying facial expressions   Aging (Excluding the usage of childhood pictures to search for grown adults)  Changing hair colors and/or styles   Beard and/or mustache,   Eyeglasses or sunglasses  Headwear such as hats or scarves  Low-light camera imaging modes  Gray-scale inputs Without requiring any operations on cloud servers, the developed face recognition system is configured to run completely on the customer’s local computer. The creation and updating of personal databases have been given to customers through an easy-to-use interface, to further increase the security of personal data.

Stage of development

TURKEY

Partner sought

The partners can be SMEs, large corporations, or government agencies. The sought partner should have the means to productize the offered face recognition and identification system in their country for any means they see fit and profitable. If desired, engineering support can be provided for application-specific optimizations, additional feature development, and/or field deployment. For a license agreement, the partner will get operational implements of the face recognition and identification system, detailed documents, and necessary information about operating and deploying the system. All this is offered against a one-time upfront payment and a license-fee for each operational instance of the system. For a commercial agency agreement, the partner will be provided with the operational executable of the offered face recognition and identification system, as well as training concerning its usage, the adoption to different use cases and additional features.