Applications of Visual Identification & Verification

Abstract

The objective of the project is to develop an application that can be used in organizations, hospitals and educational institutions to keep record of daily attendances/security. The application is based on Face Recognition & Veri cation techniques using computer vision & machine learning. The algorithm learns using the faces captures daily in the system. Another advantage is that face recognition works well for both: images of single person or a group. There are three major components in this project: 1.Client Application (Desktop and Mobile to capture/upload images), 2.Server (Cloud where actual face detection/recognition algorithms reside) and 3.Web interface (to display result statuses). There is no need for any external hardware installation. Only a device with a camera (laptop with webcam, pc with webcam, smart phone) is required for capturing/uploading images to server. The application's newer version is going to be launched with the new architecture containing the improved algorithm, APIs, and a new web interafce using the APIs. The new algorithm is made using improved layers in the neural network, classi ers and veri ers. The APIs are created to removed the tight coupling between database and server components which was existing in the older architecture, as well as third party APIs will help to create custom applications for enterprises.