An Analysis of the Spatial Layout of Housing on the Presentation of Characters' Inner Conflicts and Social Contexts in Film Narratives Based on Image Processing

Abstract

The study explores how the spatial layout of housing therein has an impact on character conflict and social context from the perspective of movie narrative context. Image processing technology is used to automatically analyze the movie frames, and a generative adversarial network model DAE-GAN based on dynamic hierarchical semantic perception is constructed, which can automatically capture spatial features from the movie frames. The experiment proves the excellent performance of DAE-GAN, which can accurately label 179 good keywords out of 385 keywords, with a 68.28% checking completeness rate and 82.55% checking accuracy rate. Especially for words with clear structure, the check accuracy rate reaches more than 94%. For the abstract atmosphere words such as “dim” and “small”, the recognition effect is also much better than other models. The empirical analysis based on 200 subjects shows that the layout of housing significantly affects the narrative, with cramped space being the most likely to trigger character conflict, with a rating of 4.24±1.14, and the social background being more prominent, with a rating of 4.04±1.17, because crowded environments naturally bring out the issues of class and resources, followed by complex spaces, while spacious and conventional layouts are relatively bland. ANOVA showed that the effects of cramped and complex layout were significant, but open space did not have a significant effect on either.