University of Hawai‘i at Mānoa – MS Plan A Thesis (2023)
Download Full Thesis (PDF)
In the realm of artificial intelligence, Automated Emotion Recognition (AER) has emerged as a pivotal area intersecting computer vision, NLP, and HCI. This thesis focuses on improving Facial Expression Recognition (FER), a core part of AER, through personalized modeling and emotional reaction intensity estimation.
Leveraging deep learning techniques—specifically CNNs, LSTMs, and Transformers—this work explores a novel suite of models designed to improve emotion classification. In particular, the study introduces a personalized CNN-LSTM approach and a CNN-Transformer method for emotional reaction intensity estimation.
Key contributions include:
This thesis bridges technical innovation with clinical potential—precisely the intersection that defines my PhD research interests. It directly informs my continued work in:
This thesis was written and implemented solely by me under the guidance of a faculty committee. It was formally reviewed, defended, and archived through the Graduate Division at UH Mānoa.