Automatic Semantic Style Transfer using Deep Convolutional Neural Networks and Soft Masks
This addresses the issue of semantic misalignment in style transfer for image synthesis, offering an incremental improvement over existing methods.
The paper tackles the problem of inappropriate texture and color application in neural style transfer by proposing an automatic method that uses semantic segmentation and soft masks to preserve content structure while transferring style, showing results that outperform recent techniques.
This paper presents an automatic image synthesis method to transfer the style of an example image to a content image. When standard neural style transfer approaches are used, the textures and colours in different semantic regions of the style image are often applied inappropriately to the content image, ignoring its semantic layout, and ruining the transfer result. In order to reduce or avoid such effects, we propose a novel method based on automatically segmenting the objects and extracting their soft semantic masks from the style and content images, in order to preserve the structure of the content image while having the style transferred. Each soft mask of the style image represents a specific part of the style image, corresponding to the soft mask of the content image with the same semantics. Both the soft masks and source images are provided as multichannel input to an augmented deep CNN framework for style transfer which incorporates a generative Markov random field (MRF) model. Results on various images show that our method outperforms the most recent techniques.