FaceCognize: An approach to Face Recognition for low resolution image
Naeem Patel, Yashveer Singh Sohi, Sai Reddy · 2021 Asian Conference on Innovation in Technology (ASIANCON) · 2021
In recent times, with the rise of the digital surveillance system, there is abundant digital forensics in the form of CCTV footage which can be utilized for the task of recognizing individuals. But due to the image resolution and environmental conditions, most of the data is rendered futile. Our proposed system is specifically aimed at efficiently leveraging this low-quality data and utilize this data from CCTV or other sources to track the individual. This system comprises of up scaling the low-quality images, extracting and aligning the faces, converting them into a feature vector, and finally using linear separators to distinguish them. This system can not only play a very important role in tracking criminal for law enforcement but also have wide-ranging application in access control, security, finding a missing person, and tracking attendance. This paper was developed as part of the Cisco ThinQbator 2020 program which is a platform to share ideas and collaborate.