Paper Title
Helmet Detection and Number Plate Recognition Using AWS Rekognition Service

Abstract
Motorcyclist safety on the road is a severe issue for traffic officials. One of the most crucial safety measures to take in the case of an accident is to wear a helmet. However, we have observed that during this epidemic, individuals use masks but fail to wear a helmet to prevent congestion, which piqued our interest, and we decided to work on a project where these helmet-less people may be fined for breaching traffic regulations. Moreover, number plate recognition is a valuable tool for law enforcement in identifying traffic offenses and stolen automobiles. To develop a reliable detection model, we made use of the AWS Rekognition service model. This object detection model uses a method known as called Faster R-CNN (Region-based Convolutional Neural Network) to identify objects in images. When AWS Rekognition service detects a helmet in an image, it returns a bounding box that surrounds the helmet, which provides information about the location and proportion of the helmet in the image compared to the bounding box of the rider. The bounding boxes for the helmet and rider are separate entities, and they are not necessarily related to each other. However, the service uses the information from both bounding boxes to make a decision about the status of helmet usage by the rider. This project proposes using picture or video footage of motorcyclists without helmets to automatically find and gather motorbike license plate numbers using Easy OCR. Keywords - Helmet Detection, AWS Rekognition Service, R- CNN, Easy OCR