Emotion Detection for Interview Analysis using CNN

Abstract

The candidate interview is a crucial part of the hiring process. Despite excelling in the preliminary rounds,
many individuals struggle to make their way through the one-on-one interview sessions. The very reason
is the lack of self-analysis, on facial expressions and confidence levels one puts forth during the interviews.
Several mock interviews evaluate the technical, verbal and logical skills of the candidates but there are
not enough means to help them to handle the final face to face interviews. The aim of this work is to
recognize and analyze the emotions displayed by the candidates using Convolution Neural Networks
(CNN) to detect the confidence levels of a person. To improve the precision of the project, eye blink rate
calculation to detect anxiety and eye gaze tracking to detect distraction are employed. The results are
consolidated and presented to the candidate as a report which can effectively assist the interview
candidates to prepare for their one on one interview preparations