Interview Performance Analysis using Emotion Detection

This project aims to assess a candidate's performance during a virtual interview process, through the use of deep learning. The method proposed involves an emotion detection module that classifies emotion based on images and nervousness from blink rate, to efficiently assess a candidate's performance. The interviewer could use this as a differentiating factor between candidates, to ensure that they recruit the best candidates for their organization. The custom model discussed performs better than the pre-trained models compared and the blink rate contributes to increased efficiency of the system. The work is done on Google Colab with Python 3 and OpenCV libraries.