Demystifying Neural Style Transfer
This provides a theoretical foundation for neural style transfer, connecting it to domain adaptation research, though it is incremental in nature.
The paper tackles the unclear principle of why Gram matrices represent style in neural style transfer by interpreting it as a domain adaptation problem, showing that matching Gram matrices is equivalent to minimizing Maximum Mean Discrepancy with a second-order polynomial kernel and achieving appealing results with other distribution alignment methods.
Neural Style Transfer has recently demonstrated very exciting results which catches eyes in both academia and industry. Despite the amazing results, the principle of neural style transfer, especially why the Gram matrices could represent style remains unclear. In this paper, we propose a novel interpretation of neural style transfer by treating it as a domain adaptation problem. Specifically, we theoretically show that matching the Gram matrices of feature maps is equivalent to minimize the Maximum Mean Discrepancy (MMD) with the second order polynomial kernel. Thus, we argue that the essence of neural style transfer is to match the feature distributions between the style images and the generated images. To further support our standpoint, we experiment with several other distribution alignment methods, and achieve appealing results. We believe this novel interpretation connects these two important research fields, and could enlighten future researches.