Image Classification Using Deep Neural Networks: Emotion Detection Using Facial Images

Introduction

Emotions enable a form of communication among human beings. Complex social communication come into context with the understanding of Emotion. Emotion Detection can be done via voice, body gestures and other complex methods. There are many practical methods to examine facial emotions too. There are seven types of human emotions that are universally recognized. The seven we are talking about includes happiness, sadness, fear, surprised, anger, disgust and neutral. A service that detects emotion from facial emotions would be widely applicable, as such a service can bring advancement in various applications of gaming, marketing, consumer product satisfaction and entertainment. Emotion Detection has attracted significant attention in the advancement of human behavior and machine learning. Various applications related to face and emotion detection include: Personal Credentials and Access Control, Video Phone and Tele conferencing, Medica land Forensic Applications, Gaming and Applications, analyzing human behavior to ascertain work satisfaction (Franco & Treves, 2001).

Face recognition is a method of identifying or verifying the identity of an individual using their face. There are various algorithms that can do face recognition but their accuracy might vary. Here I am going to describe how we do face recognition using deep learning. In computer vision, one essential problem we are trying to figure out is to automatically detect objects in an image without human intervention. Face detection can be thought of as such a problem where we detect human faces in an image. There may be slight differences in the faces of humans but overall, it is safe to say that there are certain features that are associated with all the human faces. There are various face detection algorithms but Viola-Jones Algorithm is one of the oldest methods that is also used today. Face detection is usually the first step towards many face-related technologies, such as face recognition or verification. However, face detection can have very useful applications. The most successful application of face detection would probably be photo taking. When you take a photo of your friends, the face detection algorithm built into your digital camera detects where the faces are and adjusts the focus accordingly.

There are four categories of methods that are used to detect human face, namely: Feature Method: Based on facial features like placement of eyes, nose, contour; Knowledge Method: Pre-Trained Models as we instill our model with datasets; Appearance Method: Based on Neural Networks approach; Template Method: Checks for the correlation between standard image and input image face pattern. These methods if used separately, cannot solve all the problems of face detection like orientation, pose and expression. The difficulty one face with the emotion recognition is basically due to the following: There is a moderate size database for training images, based on input image whether it is a static frame or it is a transition frame it becomes difficult to classify the emotion (Saaidia et al., 2014).

For real-time detection in which facial expressions differs dynamically, it is very difficult to detect emotions. In the present emotion detection application examine static images of facial emotions. We will inspect the system that will do the emotion recognition in real time with a live video streaming. Computation of frame-by-frame classification is necessary for live detection. So, we have developed a system for detecting emotions in real time. The result we achieved is an innovative system where an emotion-indicating text is displayed on screen.