RL-I2IT: Image-to-Image Translation with Deep Reinforcement Learning
arXiv (Cornell University)(2023)
Abstract
Most existing Image-to-Image Translation (I2IT) methods generate images in a
single run of a deep learning (DL) model. However, designing such a single-step
model is always challenging, requiring a huge number of parameters and easily
falling into bad global minimums and overfitting. In this work, we reformulate
I2IT as a step-wise decision-making problem via deep reinforcement learning
(DRL) and propose a novel framework that performs RL-based I2IT (RL-I2IT). The
key feature in the RL-I2IT framework is to decompose a monolithic learning
process into small steps with a lightweight model to progressively transform a
source image successively to a target image. Considering that it is challenging
to handle high dimensional continuous state and action spaces in the
conventional RL framework, we introduce meta policy with a new concept Plan to
the standard Actor-Critic model, which is of a lower dimension than the
original image and can facilitate the actor to generate a tractable high
dimensional action. In the RL-I2IT framework, we also employ a task-specific
auxiliary learning strategy to stabilize the training process and improve the
performance of the corresponding task. Experiments on several I2IT tasks
demonstrate the effectiveness and robustness of the proposed method when facing
high-dimensional continuous action space problems. Our implementation of the
RL-I2IT framework is available at
https://github.com/Algolzw/SPAC-Deformable-Registration.
MoreTranslated text
Key words
translation,learning,image-to-image
AI Read Science
Must-Reading Tree
Example
![](https://originalfileserver.aminer.cn/sys/aminer/pubs/mrt_preview.jpeg)
Generate MRT to find the research sequence of this paper
Chat Paper
Summary is being generated by the instructions you defined