Emotion detection

About

Facial expressions analysis plays a significant role for human computer interaction. In recent years, face emotion recognition is most important application of computer vision that can be used for security, entertainment and human machine interface. Classification of emotional features depends upon gesture, pose, facial expression, speech and etc. In this project a fuzzy method was implemented to detect emotion. The technique involves extracting mathematical data from some special regions of the face. The extracted mathematical data are then fed to a fuzzy rule based system along with the signals received from brain. These additional signals that are fed along mathematical data provide a better prediction of emotions.

Flowchart of project.

Features

Face detection

• The Viola–Jones object detection framework is the first object detection framework to provide competitive object detection rates in real-time
• Deduce emotion
• Points location method is used for feature extraction
• Finding Euclidean distance between different points on the face. This forms a basic graph based construction of the face
• As facial expressions vary, the graph changes and based on this change, we can map it to correct emotions
• Accurate mapping of emotions is achieved using fuzzy logic

Results

The accuracy rates are 94.19% for feature extraction part, 89% for standalone fuzzy logic systems and 81.8% for the complete merged system