DETERMINATION OF THE CHARACTER OF INTERVIEWEES USING TIME-SERIES CONVOLUTIONAL NEURAL NETWORKS

Recently, due to the COVID-19 pandemic, there has been an increase in non-face-to-face recruitment. Therefore, it is essential to identify the interviewee’s character in the online interview process, but there is a lack of research on analyzing the interviewer’s character using video information. In this study, we propose a deep learning model to classify personality types using landmark data of interviewees in videos. We extract the facial contour, gaze, and upper body movement coordinates of interviewees and use deep learning models to analyze features of interviewees’ personality types. In this study, a comparison experiment using real online interview videos has been conducted. As a result, Temporal Convolutional Networks (TCNs) showed the best results. The proposed model can be utilized to facilitate the company’s online interview process.