Facial Expression Classification for Online Interview Using Convolutional Neural Network

Recently, online interview plays an important role for job recruitment. Online job interviews are convenient and time-saving for interviewees. Facial expressions in the job interview provide the impression or interviewers. In this paper, we aims to classify the facial expressions from the open dataset into seven basic emotions such as Angry, Disgust, Fear, Happy, Neutral, Sad and Surprise. Online interview concentrates on two facial expressions including Happy and Surprise. The CNN model is applied to train the FER2013 datasets and improve the accuracy for the specific expressions. The experimental results demonstrate the performance of our approach by offering a robust and reliable method for facial expression classification using Convolutional Neural Network (CNN). This approach can classify the specific facial expressions with precision 83 percent and recall 84 percent for Happy and satisfy Surprise with precision 75 percent and recall 81 percent. This approach can maintain the average accuracy at 64.1 percent.