Controlling Neural Style Transfer with Deep Reinforcement Learning
This provides a user-easily-controlled style-transfer method for image processing applications, but it is incremental as it builds on existing NST techniques.
The paper tackles the problem of controlling stylization degree in Neural Style Transfer by proposing a deep Reinforcement Learning-based architecture that splits the process into steps, preserving content details early and adding style later, resulting in a lightweight method with lower computational complexity than existing one-step models.
Controlling the degree of stylization in the Neural Style Transfer (NST) is a little tricky since it usually needs hand-engineering on hyper-parameters. In this paper, we propose the first deep Reinforcement Learning (RL) based architecture that splits one-step style transfer into a step-wise process for the NST task. Our RL-based method tends to preserve more details and structures of the content image in early steps, and synthesize more style patterns in later steps. It is a user-easily-controlled style-transfer method. Additionally, as our RL-based model performs the stylization progressively, it is lightweight and has lower computational complexity than existing one-step Deep Learning (DL) based models. Experimental results demonstrate the effectiveness and robustness of our method.