HYPERPLANE CLASSIFICATION METHOD FOR IDENTIFICATION OF MIMIC MANIFESTATIONS OF EMOTIONAL STATES

Abstract

The present study aims to investigate the use of hyperplane classification to identify mimic manifestations of emotional states. Emotions are complex psychological phenomena that involve physiological, cognitive, and behavioral components. The ability to accurately recognize emotions can be compromised in certain situations, such as when individuals try to hide or fake their emotions. This is particularly relevant in security, where detecting deception and identifying emotions are critical for decision-making. Hyperplane classification is a machine-learning technique that can be used for pattern recognition and classification tasks. It involves defining a decision boundary in a high-dimensional space that separates data points into different classes. In emotion recognition, hyperplane classification can be used to identify patterns of facial expressions and other nonverbal cues associated with specific emotional states. The present study used a dataset of images depicting actors portraying different emotional states, including fear, happiness, and neutrality. The study results showed that hyperplane classification could be effective for identifying mimic manifestations of emotional states. The best-performing model was a hyperplane classification, which achieved an accuracy of 90.86% on the test set. In addition to the classification task, the authors also investigated the interpretability of the hyperplane classification models based on visual analytics. This allowed for identifying the specific facial features and movements the models used to make their predictions. The authors also discussed the limitations of the models in terms of their generalizability and robustness to variations in lighting, camera angles, and other factors. Overall, the present study proves that the proposed method of hyperplane classification for the identification of mimic manifestations of emotional states is an effective tool for identifying mimic manifestations of emotional states. Nonetheless, future research should continue to explore the interpretability and generalizability of such an approach across different populations and contexts.