Deep Sylvester Posterior Inference for Adaptive Compressed Sensing in Ultrasound Imaging
This work addresses the need for faster and more efficient ultrasound imaging in medical diagnostics, representing an incremental advance by combining adaptive sampling with deep learning for real-time applications.
The paper tackled the problem of improving ultrasound imaging frame rates and efficiency by minimizing scan-lines, introducing an adaptive subsampling method that uses Bayesian posterior inference and deep generative models to maximize information gain, resulting in a 15% improvement in mean absolute reconstruction error and real-time performance at 66Hz.
Ultrasound images are commonly formed by sequential acquisition of beam-steered scan-lines. Minimizing the number of required scan-lines can significantly enhance frame rate, field of view, energy efficiency, and data transfer speeds. Existing approaches typically use static subsampling schemes in combination with sparsity-based or, more recently, deep-learning-based recovery. In this work, we introduce an adaptive subsampling method that maximizes intrinsic information gain in-situ, employing a Sylvester Normalizing Flow encoder to infer an approximate Bayesian posterior under partial observation in real-time. Using the Bayesian posterior and a deep generative model for future observations, we determine the subsampling scheme that maximizes the mutual information between the subsampled observations, and the next frame of the video. We evaluate our approach using the EchoNet cardiac ultrasound video dataset and demonstrate that our active sampling method outperforms competitive baselines, including uniform and variable-density random sampling, as well as equidistantly spaced scan-lines, improving mean absolute reconstruction error by 15%. Moreover, posterior inference and the sampling scheme generation are performed in just 0.015 seconds (66Hz), making it fast enough for real-time 2D ultrasound imaging applications.