Photorealistic Style Transfer via Wavelet Transforms
This work improves style transfer for applications requiring photorealism, such as photography and video editing, by introducing a novel correction to network architecture.
The paper tackles the problem of photorealistic style transfer by addressing spatial distortions and unrealistic artifacts in existing methods, achieving a model that stylizes 1024x1024 resolution images in 4.7 seconds with pleasing quality and stable video stylization.
Recent style transfer models have provided promising artistic results. However, given a photograph as a reference style, existing methods are limited by spatial distortions or unrealistic artifacts, which should not happen in real photographs. We introduce a theoretically sound correction to the network architecture that remarkably enhances photorealism and faithfully transfers the style. The key ingredient of our method is wavelet transforms that naturally fits in deep networks. We propose a wavelet corrected transfer based on whitening and coloring transforms (WCT$^2$) that allows features to preserve their structural information and statistical properties of VGG feature space during stylization. This is the first and the only end-to-end model that can stylize a $1024\times1024$ resolution image in 4.7 seconds, giving a pleasing and photorealistic quality without any post-processing. Last but not least, our model provides a stable video stylization without temporal constraints. Our code, generated images, and pre-trained models are all available at https://github.com/ClovaAI/WCT2.