Table 1.

Different regions (single/pair) of the face, with the most discriminative facial landmarks. Single region (Ri) means landmarks within this region are used to build a feature vector either from the stand alone landmark tracking result or pair of landmark tracking result. Pair of region (Ri–Rj) means two landmarks, one from each, is used to build the feature vector from the tracking result of those landmarks.

Expression	Face Regions (Single/Pair) with the Most Discriminative Landmarks
Anger	R1-R2, R2, R2-R3, R5-R7, R6, R6-R7
Disgust	R1-R2, R2, R2-R3, R5, R5-R6, R5-R7
Fear	R1-R2, R2, R3-R6, R5-R6, R2-R7, R4-R7, R6-R7
Happiness	R1-R2, R2, R2-R4, R2-R6, R3-R6, R4-R6
Sadness	R2-R3, R2, R3-R7, R5-R6, R6, R5-R7, R6-R7
Surprise	R1-R3, R2, R2-R3, R2-R7, R2-R5, R7