查看文章

There has been a recent shift from using text-based sentiment analysis in favor of an image-based method. In recent years, transfer learning methods have been widely utilized in developing a comprehensive image sentiment analysis approach. Deep learning algorithms have produced remarkable outcomes in a variety of contexts. Image-based sentiment analysis presents many difficulties, but there also appears to be much space for development. A significant improvement over prior work is provided by an InceptionV3 approach that can easily focus on huge body portions like a human face. This research improves image categorization performance with InceptionV3, a popular deep convolutional neural network, and other deep features. Using a Convolutional Neural Network based on InceptionV3 architecture, we identify and classify emotions using the famous CK + , FER2013, and JAFFE datasets …