Human Traits Classification/Detection Using Deep Learning

Affective computing, intelligent medicine, personalized recommendation analysis, and other fields have shown a strong interest in automated human trait classification and detection from any available human data, including text, image/video, and voice. This research provides a detailed and comprehensive study of human traits such as movement of hand and iris during the talk, human physical characteristics as well as the beard, mustache, face shape types, and eye color, together with emotions from audio for scientific and psychological analysis. Each human trait is analyzed, and specific algorithms are implemented to meet the requirements using computer vision and deep learning methods. This study has been mainly based on neural networks, specifically deep learning models and techniques. To reach the goals, novel methodologies have been implemented for this study, including human traits classification and detection, under the supervision of a company. Moreover, this study has taken advantage of creating a new dataset for one specific task to overcome the state-of-the-art results to implement it for a real-world scenario. The creation of the dataset includes collecting relevant images under certain conditions from various people/environments and annotating them based on the usage requirements.