Multimodal Transfer: A Hierarchical Deep Convolutional Neural Network for Fast Artistic Style Transfer
This work addresses a specific issue in style transfer for high-resolution images, offering incremental improvements in texture fidelity and scale handling for applications in digital art and photography.
The authors tackled the problem of artistic style transfer on high-resolution images, where existing methods fail to capture intricate textures and maintain correct scales, by proposing a hierarchical multimodal CNN that processes color and luminance channels with multiple losses, achieving visually pleasing results more similar to desired styles in nearly real-time.
Transferring artistic styles onto everyday photographs has become an extremely popular task in both academia and industry. Recently, offline training has replaced on-line iterative optimization, enabling nearly real-time stylization. When those stylization networks are applied directly to high-resolution images, however, the style of localized regions often appears less similar to the desired artistic style. This is because the transfer process fails to capture small, intricate textures and maintain correct texture scales of the artworks. Here we propose a multimodal convolutional neural network that takes into consideration faithful representations of both color and luminance channels, and performs stylization hierarchically with multiple losses of increasing scales. Compared to state-of-the-art networks, our network can also perform style transfer in nearly real-time by conducting much more sophisticated training offline. By properly handling style and texture cues at multiple scales using several modalities, we can transfer not just large-scale, obvious style cues but also subtle, exquisite ones. That is, our scheme can generate results that are visually pleasing and more similar to multiple desired artistic styles with color and texture cues at multiple scales.