The Application of Emotion Recognition Based on Deep Learning in Photography Works

In order to understand the application of emotion recognition in photography, research on the application of emotion recognition based on deep learning in photography works has been proposed. This article first proposes an improved Word2Vec method for constructing an emotional dictionary for photography works. Secondly, the importance of the Word2Vec training result set is modified using the TF-IDF training result word set and the artificial photography emotion basic word set to generate the photography work emotion word set. Seed words and emotion word sets are selected from the university emotion vocabulary ontology library to calculate the polarity and strength of emotion words, and combined with the HowNet emotion dictionary to form a photography work emotion dictionary. Finally, the experimental results indicate that this method has achieved good results on actual photography datasets.