DiffStyler: Diffusion-based Localized Image Style Transfer
DiffStyler introduces a novel approach for efficient and precise arbitrary image style transfer, surpassing previous methods in achieving a harmonious balance between content preservation and style integration.
Mamba-ST: An Efficient State Space Model for Image Style Transfer
Mamba-ST, a novel state space model architecture, can efficiently perform image style transfer by adapting the inner equations of Mamba to fuse content and style information without requiring additional modules like cross-attention or custom normalization layers.
PixelShuffler: A Simple and Efficient Approach to Image-to-Image Translation for Style Transfer
This paper introduces PixelShuffler, a novel and efficient image style transfer method that leverages pixel shuffling guided by mutual information maximization to effectively combine the content of one image with the style of another while preserving structural details and bypassing the need for complex neural network architectures.
© 2024 by Linnk AI