Reinforcement Learning for Sampling on Temporal Medical Imaging Sequences

TL;DR

This paper demonstrates that reinforcement learning algorithms, specifically double deep Q-learning and REINFORCE, can effectively learn optimal sampling strategies for dynamic medical imaging sequences, improving reconstruction quality with fewer measurements.

Contribution

It introduces a novel application of reinforcement learning to optimize sampling patterns in temporal medical imaging, leveraging a pre-trained autoencoder for reconstruction.

Findings

RL algorithms successfully learn effective sampling patterns.

Optimal sampling strategies improve image reconstruction quality.

Proof of concept validated on medical imaging data.

Abstract

Accelerated magnetic resonance imaging resorts to either Fourier-domain subsampling or better reconstruction algorithms to deal with fewer measurements while still generating medical images of high quality. Determining the optimal sampling strategy given a fixed reconstruction protocol often has combinatorial complexity. In this work, we apply double deep Q-learning and REINFORCE algorithms to learn the sampling strategy for dynamic image reconstruction. We consider the data in the format of time series, and the reconstruction method is a pre-trained autoencoder-typed neural network. We present a proof of concept that reinforcement learning algorithms are effective to discover the optimal sampling pattern which underlies the pre-trained reconstructor network (i.e., the dynamics in the environment). The code for replicating experiments can be found at https://github.com/zhishenhuang/RLsamp.