Abstract:
Existing attendance systems are entirely manual or semi-automatic. They are much time consuming and difficult to manage. So our aim was to build a system which is truly automated, marks attendance in no time and is self-managed. To achieve our goals we decided to build a facial recognition based automated attendance system. So our project emphasises on using a modern machine learning techniques to perform highly accurate face recognition. We have used deep convolutional neural networks to achieve the desired accuracy. Further we developed an extensive application that uses our face recognition model to mark attendance in classes and work environments and is based on a scalable architecture. Comparing our application with existing systems, it is much efficient in terms of memory and processing. In future our aim is to improve accuracy of our facial recognition module so that it can give better results in different lighting-conditions.
