A Robust Design of Real-Time Resilient Smile Recognition System using Hybrid Deep Learning Principles

Abstract

We often convey our happiness via smiling or making a happy face. Simple, latent feelings like contentment and joy are brought to the surface in this way. In terms of social communication, this is the most difficult and potent task there is. One function of digital cameras is smile recognition, which prevents taking pictures of people unless they are smiling. There are two distinct stages to smile detection. After identifying a face in a video or picture, it watches for a grin. To be more precise, we may state that a motion detector takes a video or image and breaks it down into frames. It then examines the frameworks, such as the balanced level and flash, for the face area. The camera can tell whether someone is laughing or smiling because it looks for certain facial features—a jutted jaw, narrowed eyes, and an elevated cheekbone among others. The components of computer architecture and software advancements are advancing at a rapid pace, which is leading to a progressive increase in customer needs for electric products. A growing number of people are interested in human computer interaction systems that differ from the conventional user interface, which includes the keyboard and mouse. This article depicts the realtime implementation of deep learning algorithms in computer settings with limited resources. In this study, we provide HLSR, a new kind of deep learning model that combines the XGBoost and Convolutional Neural Network (CNN) algorithms for smile recognition. To get a clear picture of how well the suggested technique works, we cross-validate the HLSR model with the current CNN model.