A study on Evolution of Facial Recognition Technology

Since the inception of Facial Recognition (1960s) researchers began experimenting with computer-based facial recognition algorithms, but they were incompetent due to the limited processing power of computers. Then researchers developed feature-based recognition systems in the 1980s, which identified certain facial characteristics, such as the space between the eyes or the nose’s form, etc. to create a unique facial signature, however, they were still limited in their accuracy. 3D facial recognition systems were introduced in 1990s, which used depth perception to create more accurate facial models. These systems were primarily used in security and surveillance applications. Machine learning algorithms in 2000s could learn to recognize faces more accurately over time because it uses large datasets to train themselves to recognize patterns in facial features. Deep learning algorithms of 2010s could recognize faces with even greater accuracy as they use neural networks to analyze facial features at multiple levels of abstraction, allowing them to identify complex patterns. Real-time facial recognition systems were also developed during this period to recognize faces in real-time video streams and therefore found applicable in security and marketing. Covid-19 Pandemic incorporated Facial recognition technology with facemask requiring additional considerations and adjustments in order to be effective in accurately identifying individuals who are wearing masks. This paper presents a study of evolution of Facial recognition technology as viable biometrics since its inception and how it got molded over time due to technological, legal and global interventions. At the end, we conclude this paper with promising directions for future research on this field.