Image Based Sentiment Analysis using Bayesian Networks and Deep Learning

The automated recognition of facial emotions is a challenging task that has recently attracted much logical attention. However, the topic of feeling acknowledged by a group of people has received less attention. However, the vast amount of information available on person-to-person communication websites, including images of crowds of people traveling to various get-togethers, is slowly gaining popularity. However, image-based sentiment analysis is a challenging subject due to challenges such as change in the head and body position, occlusions, other lighting conditions, actor fluctuations, various indoor and outdoor situations, and picture quality. This study tries to categorize felt emotions as Positive, Neutral, or Negative for individuals or a group of individuals. This study presents a Deep Learning model and Bayesian classifiers-based approach to picture sentiment analysis. The model's 77.58 percent accuracy is in line with the most recent findings.