Research on Sentiment Classification of Active Scene Images Based on DNN

Abstract

With the development of computer and communication technology, more and more activity scene design images have been digitized. In order to solve the problem of automatic classification of such scene images, this paper proposes a multi-featured activity scene image sentiment classification method. This research analyzed the relationship between color features (artistic style) and emotional semantics and established the activity scene image data base. According to the characteristics of the scene image. It extracted the global and local color features based on the characteristics of the scene image and completed the sentiment classification of the activity scene image with classifier of deep neural network.