Interest Indices for Human Computer Interaction Applications Based on Facial Image Sequences

Title: Interest Indices for Human Computer Interaction Applications Based on Facial Image Sequences
Author: Shubhobrata Bhattacharya{1}, Anirban Dasgupta{1}, Anwesha Sengupta{2}, Aurobinda Routray{1}
Affiliation: {1}Indian Institute of Technology Kharagpur, India; {2}Techno Main Salt Lake, India


Abstract: The present paper proposes three metrics for assessing interest levels of users in human-computer interaction (HCI) applications using facial features. The indices, named Expression Index (EI), Ocular Index (OI), and the Head Pose Index (HPI), are based respectively on the three most significant aspects of HCI, viz., facial expression, eyelid dynamics, and head pose. The proposed indices have been implemented using Convolutional Neural Networks (CNN) and validated through statistical analysis.

Head Pose,Eyelid Motion,Expression,CNN,IEEE Sensors Council,IEEE Sensors 2019,1503,