| Autor: |
Singh, Krishna Kumar, Chiranjeevi, Sadu, Sivalal, Kethavath |
| Předmět: |
|
| Zdroj: |
International Conference Computer Graphics, Visualization, Computer Vision & Image Processing; 2021, p45-52, 8p |
| Abstrakt: |
Personality assessment has been widely used in the professional psychology and signal processing fields. Recently, it has been a great interest from the computer vision research community in assessing personality from visual data. Many state-of-the-art models are assigned the Big-Five personality indicators using either external judges or personal interviews. We propose Face Features-based Personality Assessment (FFPA) that assesses the personality of a person based on one's facial features. It maps facial appearance into the Big-Five personality indicators, namely Extraversion, Agreeableness Conscientiousness, Neuroticism and Openness. The geometry-based and appearance-based approaches are used to extract features from the face and mapped to the personality indicators using Partial Least Square Regression (PLSR). The corresponding personality indicator values are collected by filling the online form of the Big-Five personality assessor. The computational experiments are performed on a synthetic dataset, consisting of the face images of 200 students. The experimental results show that the proposed model predicts the personality indicators with 0.95 coefficient of determination (approx.) and Mean Squared Error (MSE) is 0.001 (approx.). [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
