Detecting facial expressions from kinect camera

• Main Author: Muhammad Hafiz Bin Mohd Zakee
• Other Authors: Justin Dauwels
• Format: Final Year Project
• Language: English
• Published: 2016
• Subjects: DRNTU::Engineering::Electrical and electronic engineering::Control and instrumentation
• Online Access: http://hdl.handle.net/10356/68282
• Institution: Nanyang Technological University

In the past years, there were several advances methodology regarding face detection and tracking, features extraction methodology and techniques used for expression classification. Detection of facial expression is a very common communication method between computer and human interface. The facial behaviour according to emotion is an important element in human communication, as it carries an amazing amount of information that can reflect emotional feelings. Observing person’s facial expressions or behaviours assist a person understand their emotional feelings. New technology provided today for detecting facial expressions, with rapid and high resolution image acquisition, helps us to analyse and recognize in real time facial expressions. This application can be useful in many real time applications like military security, trading (the customer’s emotions about a product), patient monitoring, and others. This paper presents an application that detects three main basic facial expression (Neutral, Happy and Sad) by using Microsoft Kinect for Windows sensor V1. To detect the facial expressions, facial parameterization using Facial Action Coding System (FACS) were extracted from the recording by face tracking SDK provided by Microsoft Kinects. Four FACS trained annotators were employed to manually label the facial expressions by viewing a videotaped recording of 11 subject’s facial behaviours from Kinect Studio. A machine learning algorithm, KNN and Decision Tree will classify the facial expression sin real time.