D.E.E.P. – Dynamic Emotion Extraction Process: A Machine Learning-Based Approach for Facial Expression Analysis and Emotional Trends – Case studies with real images (Processo de extração dinâmica de emoções: Uma abordagem baseada em machine learning para análise de microexpressões faciais e tendências emocionais – Estudo de caso com imagens reais)

Resumo

https://doi.org/10.5281/zenodo.20209786

This paper presents the development of D.E.E.P. (Dynamic Emotion Extraction Process), an innovative machine learning-based computational tool. The primary goal of D.E.E.P. is to identify emotional and behavioral trends through detailed analysis of facial expressions and facial expressions extracted from videos. The comprehensive methodology encompasses visual data collection, sophisticated image processing, robust machine learning model training, and the subsequent generation of an emotional timeline. This second version of the work elaborates on the theoretical foundations supporting the solution, as well as the main challenges encountered during development, such as overfitting issues and initial difficulties with K-Means clustering. It also highlights the integration of a novel dataset obtained in collaboration with Dr. Wilma Bainbridge, Ph.D., from the University of Chicago, and the documentation of code progression. Preliminary results suggest the practical applicability and robustness of the proposed system.