| Sumario: | This research explores how pupillometry can be used to recognize emotions by studying changes in pupil size in reaction to emotional cues. Sophisticated machine learning models such as CNNs, transformer-based structures, and ensemble methods are utilized to tackle the difficulties of real-time applications. The study emphasizes the benefits of incorporating pupillometry into multimodal frameworks that incorporate other physiological signals like heart rate variability and skin conductance to enhance the reliability and precision of emotion recognition systems. The study's main contributions are the advancement of adaptive systems that can react to real-time emotions, the establishment of datasets for training and validation, and the investigation of multimodal techniques to improve performance in different scenarios. Moreover, the results highlight the possibility of using pupillometry-based models in areas like mental health, education, and human-computer interaction, where personalized and adaptive emotion recognition can make a big difference. This study sets the foundation for future progress in adaptive emotion recognition systems by tackling challenges in scalability, environmental variability, and data diversity. Combining pupillometry with advanced machine learning methods shows potential for improving emotional intelligence in technology, opening opportunities for use in virtual reality, therapy, and adaptive interfaces in real-world scenarios. This research helps advance the field of affective computing by offering a detailed structure for identifying emotions using nonintrusive physiological measures. Our contribution to this paper is to summarize current research on pupillometry for emotion recognition, emphasizing deep learning approaches, multimodal integration, and real-time application challenges. We highlight advancements, limitations, and future research directions to enhance emotion-aware systems in healthcare, education, and human-computer interaction.
|
|---|