Automated Attendance system for the Entire Classroom with Multi-Face Facial Recognition via AWS Rekognition

Taking attendance is crucial for any class or meeting to account for a person's presence on the spot. Manual attendance methods are time-consuming, disruptive, and prone to proxy cheating. Existing methods such as QR scanning, iris recognition, and fingerprint recognition take more time and create congestion during the scanning process. Modern society requires an advanced method for taking attendance that is easily implementable, accurate, faster, and non-disruptive to any program. This study presents an advanced cloud-based attendance system with face recognition using the AWS API, which can provide accurate results from a single image containing multiple faces. The system utilizes AWS Rekognition for face detection and matching, Amazon S3 for secure storage of user images, and AWS Lambda for executing logic in an environment. DynamoDB serves as a database for storing registered student IDs and face prints and is enhanced by a user-friendly front-end interface to facilitate real-time interactions. Our designed model has been tested using various real-time images and has achieved 100% accuracy. It provides the total number of detected faces, indicating how many are present and how many are absent in the image, along with the number of unknown faces and their images. The proposed system is easy to implement, faster, more cost-effective, a time saver, and hassle-free to operate.