Abstract

"Facial Expression Detection" is one of the most challenging tasks in the field of object detection. Object detection is a field of computer vision that deals with detection and recognition of semantic objects (such as cars, human face, buildings and etc) in images and videos. Nowadays, with advancements in hardware technologies and increase in image acquisition. Facial expression in human images can convey important information. Facial expression can be used for deciding marketing and teaching strategies. Facial expressions along with hand gestures can be used in VR applications. It can useful in providing inclusive healthcare services. In this paper, we propose a novel approach to extract facial features and recognize facial expression from an image. We display our hybrid approach to detect and measure facial features such as eyes, nose, mouth and eyebrows. We also describe various feature points and their associated distance which can be used for expression recognition by matching techniques using Supervised Machine Learning. We have used the JAFFE (Japanese Female Facial Expression) database for this project. We classify each image into seven categories of facial expression viz. 1) Angry, 2) Disgust, 3) Fear, 4) Happy, 5) Neutral, 6) Sad and 7) Surprise.