MediaRoad assigned the Quality Label to a Sandbox project namely Sensifai celebrity recognition system for VRT. The Sensifai’s celebrity recognition system has been integrated into VRT’s archives to empower content search. Media companies, publishers, social media and multimedia asset management providers often require finding and tagging celebrities in large archives of video and images. Powered by visual deep-learning technology, Sensifai offers an easy-to-use interface which automates the process of recognizing celebrities with a very accurate face recognition/identification system. This model can recognize more than 3500 celebrities from all over the world. The product has been released publicly in cooperation with AWS Sagemaker platform. The launch of this platform and AWS marketplace for machine learning is announced during AWS re:Invent 2018.
It allows you to scale and then search digital image and video libraries. A very initial version of the software was successfully applied to VRT video archive as a Sandbox project to find and tag celebrities of interests. After the Sandbox project with VRT the software was significantly improved and today Sensifai’s celebrity recognition system on Amazon Sagemaker recognizes over 3500 celebrities with an accuracy of 98%. This success attracted several customers.