Across India there are reports coming in about use of facial recognition systems to track people assembling to protest against the citizenship amendment act. Indian Express has reported the Delhi Police has acquired automated facial recognition system to find missing children because of an order by Delhi High court. Other police departments like Hyderabad, Chennai and Surat police have acquired facial recognition systems from different vendors. At the same time the National Crime Records Bureau has issued a tender to procure a national facial recognition system to be supplied to all police departments. This is systems will be part of the crime and criminal tracking numbering system.
Any Facial recognition software reads the geometry of a face captured from a photo or video to create a unique code or ‘faceprint’. This is essentially a set of unique virtual points on your face. These points of geometry are computed using algorithms of image processing and algorithms of machine learning are used to train with these faceprints to find similar data points from existing identity databases. These advancements in facial recognition technology were made over time, with a push from image processing requirements of driverless cars to social media pushing it to automatically tag your friends in photos. Between 2014 and 2018 facial recognition systems got 20 times better according to US National Institute of Standards & Technology.
In this episode, Srinivas Kodali speaks with Vidushi Marda, a lawyer and programme officer from article 19, where she works on artificial intelligence systems and has been focusing on state use of facial recognition systems in her research. she is also a non-resident research analyst at Carnegie India where her work focuses on facial recognition systems by police authorities in India.