Facial and posture features assisted personality traits recognition from videos

Abstract

There is a growing interest in detecting personality of an individual in a non-intrusive manner in areas like career development, counseling, interviews etc. Many solutions have been proposed for detection of personality from facial features alone. Different from it, this work proposes a model for detection of personality integrating the features learnt from face and human postures from videos.  An integration of traditional and deep learning features in combination with machine learning algorithm is used to classify the personality of the individual. This work also proposes a frame of interest selection algorithm for selection of suitable frames in the video for personality assessment.