Emotion Detection by Employing Deep Learning CNN Model

Facial expression recognition through computer algorithms presents many challenges, despite its innate human capability. This study compares some of the most renowned methods used in Emotion detection and proposes VGG-16 as the most tailored method for emotional recognition, emphasizing the importance of facial expressions as emotional indicators. Leveraging recent advancements, including Robert Horling’s cognitive processes, this integrated system adeptly discerns the subjects’ moods. While electroencephalography remains a formidable tool, the method proposed by this research paper offers a cost-effective alternative for emotion recognition. Through comprehensive comparative studies and empirical evaluations, this research demonstrates VGG-16’s superior performance in terms of both training speed and accuracy, surpassing alternative models such as ResNet-50 and AlexNet. VGG-16 achieved an accuracy rate exceeding $71 \%$, outperforming ResNet- 50 with $68 \%$ accuracy and AlexNet with approximately $64 \%$ accuracy. The results show the efficacy of VGG-16 in the context of training efficiency and accuracy, substantiating its merit as a robust model for emotion detection in comparison to alternative architectures. This research opens avenues for real-world applications in diversified sectors such as education, healthcare, and HCI (Human-computer interaction).