DEVELOPMENT OF MULTIMODAL PERSONALITY PREDICTION MODEL USING PERSONALITY TRAITS

Abstract

It might be challenging to predict someone's personality in both the workplace and in daily life. There are several variables that can affect personality prediction, and these variables can change from person to person. Personality reflects an individual’s behaviour, thought process , life choices, mental health , emotions, social character. Variou deep learning model has been used in this project for multi modal personality prediction. VGGish convolutional networks (VGGish CNN), Resnet50, InceptionV3, Xception, Convnext, InceptionResnet have been used to extract facial and ambient features from the video, 2D convolutional neural network and Alexnet, have been to extract audio features. For the final prediction the extracted feature is given as a input to a fully connected layer followed by sigmoid activation function.