An augmented reality based Virtual dressing room using Haarcascades Classifier

Virtual dressing rooms are modern way of shopping where customers experience realistic dress fitting and order online. As trying different dresses at shopping centers is very time consuming and customers are not sure how the dress looks on them. Therefore, there is a need to identify efficient methods to improve virtual dressing rooms and increase e-commerce businesses. All the existing virtual dressing room methods required Kinect sensors for virtual dress fitting and Kinect sensors are expensive. Our motivation is to create such virtual dressing room that overcome the constraint of Kinect sensor input. We present a dynamic virtual dressing room based on web cam to acquire input video and allow user to experience virtual dress fitting. Our proposed model comprised of following steps including face, skin, lower body detection and dynamic distance estimation for precise virtual reality. We have evaluated our algorithm on 50 subjects using 10 dresses.