Human emotion detection through facial expressions for commercial analysis

Emotion defines what do we feel about something and facial expressions are the most obvious way to show it. Identifying human emotions through the use of a computer has always been a challenge thus, Facial Expression Recognition (FER) systems are being developed to mainly detect and recognize seven (7) core emotions namely: Joy, Fear, Sadness, Neutral, Surprise, Disgust and Anger. The proposed work introduces an Artificial Neural Network (ANN) which models the relationship between human facial expression and corresponding conveyed emotion extracting visual features while viewing a commercial. ANN also classifies the image set into the seven core emotions. Dimensionality Reduction based on Principal Component Analysis (PCA) was done to obtain feature vectors of the image dataset. Improved recognition rate can be seen on the experimental data which is due to the dimensionality reduction performed. The proposed algorithm is tested using the Japanese Female Facial Expression (JAFFE) Database.