Deep Scatter Estimation (DSE): Accurate Real-Time Scatter Estimation for X-Ray CT Using a Deep Convolutional Neural Network

Abstract

X-ray scatter is a major cause of image quality degradation in dimensional CT. Especially, in case of highly attenuating components scatter-to-primary ratios may easily be higher than 1. The corresponding artifacts which appear as cupping or dark streaks in the CT reconstruction may impair a metrological assessment. Therefore, an appropriate scatter correction is crucial. Thereby, the gold standard is to predict the scatter distribution using a Monte Carlo (MC) code and subtract the corresponding scatter estimate from the measured raw data. MC, however, is too slow to be used routinely. To correct for scatter in real-time, we developed the deep scatter estimation (DSE). It uses a deep convolutional neural network which is trained to reproduce the output of MC simulations using only the acquired projection data as input. Once trained, DSE can be applied in real-time. The present study demonstrates the potential of the proposed approach using simulations and measurements. In both cases the DSE yields highly accurate scatter estimates that differ by < 3% from our MC scatter predictions. Further, DSE clearly outperforms kernel-based scatter estimation techniques and hybrid approaches, as they are in use today.

Real‐time scatter estimation for medical CT using the deep scatter estimation: Method and robustness analysis with respect to different anatomies, dose levels, tube voltages, and data truncation

“…Therefore, we recently proposed the deep scatter estimation (DSE) for medical CT and industrial CT . To estimate scatter, DSE uses a deep convolutional neural network which is trained to reproduce the output of MC simulations given only a function of the acquired projection data as input.…”

Projection‐domain scatter correction for cone beam computed tomography using a residual convolutional neural network

“…The scatter‐only or scatter‐free projections are not attainable in CBCT experiments. Available solutions are to use projections calculated by pre‐evaluated MC simulations, to use higher quality images like fan‐beam CT images or to apply image‐to‐image translation networks with unsupervised learning . Our MC simulation setups were referenced from specifications of a real CBCT machine, but the FPD grid was not considered because the data were not available.…”

“…Specific to medical imaging, various CNN models have been developed for low‐dose fan‐beam CT image restoration, MR‐to‐CT image symthesis, PET image segmentation, and so on . Recently, Maier et al proposed a CNN‐based scatter correction for industrial CBCT applications, and Hansen et al demonstrated that CNN‐based CBCT intensity correction improved photon dose distribution calculation accuracy.…”

Registration-based multi-orientation tomography

“…To improve the quality of CT reconstruction based on sub-optimal data, an ever increasing number of advanced computational approaches are being developed. Such methods include: model based and empirical beam hardening corrections [9][10][11][12], statistical noise models [13], regularized reconstructions [14], machine learning-based approaches [15,16] and approaches based on hardware calibration [17].…”