Kartikeya Bhardwaj, a Mechanical Engineering student (class of 2020) of Mahindra University, Hyderabad has created a deep learning app, Chakshu Face Attendance, to provide employers/institutions/organisations with touchless recording of attendance attributes. Especially relevant in today’s Covid outbreak, this app provides a safe alternative to touch-based fingerprint attendance systems. The start-up is being incubated at Mahindra e-hub, an incubation centre of Mahindra University under the University’s Centre of Entrepreneurship & Innovation.

“Mahindra University focuses on providing an enabling atmosphere to the students so that they can apply their learning in creating innovative solutions to everyday problems. One such focus has been on Artificial Intelligence and Deep Learning, wherein we have not just created state-of-the-art labs but also made available live projects for students to work on. It is indeed exciting to witness these efforts coming to fruition and our Centre for Entrepreneurship and Innovation is providing a platform for these start-ups to grow and grow on their own by providing all necessary assistance in sync with the Indian Government’s mission”, says Dr. Yajulu Medury, vice chancellor, Mahindra University.

In big organisations, for attendance, fingerprint or ID scanners are the norm. In Covid times, “touchless” technology is the call of the hour since anything that is used by many people may turn out to be conduit for transmission of the virus. In MSMEs, this isn’t so, primarily because of the issue of costs, e.g. RFID systems require expensive hardware and regular maintenance. In addition, a common issue is that of Buddy Punching, where employees do proxy attendance of their colleagues using their IDs results in huge losses and is an unethical practice that still floats around in offices all over the world.

Chakshu.ai has developed an app named Chakshu Face Attendance, a Touchless Employee Attendance App that lets an organisation Clock attendance through an android device (phone or a tablet), using Face Recognition. The Attendance process is Completely Touch-Free, and Covid-19 ready. It also eliminates the Buddy Punching problem through SpoofSense: Chakshu’s AI algorithm is trained to detect Spoofing attacks and reject them, thereby making the process secure and trustable.

Chakshu Face Attendance App uses PyTorch (a Powerful Machine Learning Framework by Facebook) to build and deploy Deep learning algorithms. Scaling AI algorithms to thousands of users is a compute intensive process and often results in high latency situations. Chakshu App involves careful model optimisations that makes the backend capable to serve a huge volume of traffic at peak performance, ensuring a frictionless user experience.

“A strong anti spoofing provision makes Chakshu Face Attendance a highly reliable face attendance system. The app is secure, efficient and highly scalable”, says Kartikeya Bhardwaj, founder, Chakshu.ai.