Comic Character Personality Prediction Based on Multimodal Fusion

The appearance and lines of comic characters can reflect the personality of comic characters and help improve comic works. Previous deep learning-related work in the field of comics has rarely involved character analysis in comics, and the mode of feature extraction is single. In this paper, a multimodal personality reasoning method is proposed, which uses the visual features of comic characters and the text features of lines to predict the scores of the Big Five personality traits by combining the image features and text features. The effectiveness of the proposed method is verified through experiments, and the multimodal fusion method is superior to each single modal channel, which shows the potential of comic character reasoning, and provides a new way of thinking and method for personality analysis based on deep learning.