Analyzing and evaluating personality and human behavior based on facial index and big five model
Year: 2024
Type: article
Abstract: Numerous studies have shown that morphological and social indicators in a human face can provide information about a person's personality and behavior. The Big Five model, also known as the Five-Factor, is the five basic dimensions of personality. These dimensions include openness, conscientiousness, extraversion, agreeableness and neuroticism. The Big Five model has been applied in a variety of different settings, including clinical psychology, organizational psychology, and even marketing research. By examining where an individual falls on each of these dimensions, researchers can gain insight into their unique personality traits and use this information to make predictions about their behavior and performance in different situations. In our existing iscv platform, a job searching website, we can help employers better understanding employee incentives by utilizing the personality traits information of candidates. Managers and CEOs can therefore discover a means to improve relationships and communication while also managing and building teams more effectively. We trained a machine learning model using a hybrid CNN-LSTM, ResNet, VGG19 algorithm for personality recognition through interview video. In each video, we analyze facial movement by using the 3D landmarks extracted with the 3DDFA-V2 algorithm. The model uses the UDIVA v0.5 dataset, collected in the scope of the research project entitled “Understanding Face-to-Face Dyadic Interactions through Social Signal Processing”. The experimental results conclude: (i) Analyzing facial movement by using the 3D landmarks extracted with the 3DDFA-V2 algorithm. (ii) Personality traits inferred from facial behaviors by most benchmarked deep learning model. (iii) Personality assessment model is trained from a combination of two data sets (one UDIVIA dataset and one self-survey dataset) to fit Asian personalities. (iv) The detailed Big Five personality tendency assessment table is based on the interview video and questionnaire of the surveyed people.
Authors: Dinh Thuan Nguyen, Minh Nhut Nguyen, Anh Thu Le, Nam Do Dang Kien, Quan Dang Minh
Language: English
Topic: Emotion and Mood Recognition
Subfield: Experimental and Cognitive Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: diamond
Award: G5688350078
OpenAlex ID: W4409521943