Deep Learning Driven Noise Reduction for Reduced Flux Computed Tomography

Abstract

Deep neural networks have received considerable attention in clinical imaging, particularly with respect to the reduction of radiation risk. Lowering the radiation dose by reducing the photon flux inevitably results in the degradation of the scanned image quality. Thus, researchers have sought to exploit deep convolutional neural networks (DCNNs) to map low-quality, low-dose images to higher-dose, higher-quality images, thereby minimizing the associated radiation hazard. Conversely, computed tomography (CT) measurements of geomaterials are not limited by the radiation dose. In contrast to the human body, however, geomaterials may be comprised of high-density constituents causing increased attenuation of the X-rays. Consequently, higher-dose images are required to obtain an acceptable scan quality. The problem of prolonged acquisition times is particularly severe for micro-CT based scanning technologies. Depending on the sample size and exposure time settings, a single scan may require several hours to complete. This is of particular concern if phenomena with an exponential temperature dependency are to be elucidated. A process may happen too fast to be adequately captured by CT scanning. To address the aforementioned issues, we apply DCNNs to improve the quality of rock CT images and reduce exposure times by more than 60%, simultaneously. We highlight current results based on micro-CT derived datasets and apply transfer learning to improve DCNN results without increasing training time. The approach is applicable to any computed tomography technology. Furthermore, we contrast the performance of the DCNN trained by minimizing different loss functions such as mean squared error and structural similarity index.

Generation of training dataset for deep-learning noise reduction

“…14 Deep learning (DL) has been explored in recent years for CT noise-reduction as part of the image reconstruction or the post-processing. [14][15][16][17][18][19][20][21] When proper training datasets are used, DL has shown significant noise-reduction capability while preserving the FBP-like peak-frequency in NPS. In DL, the production of the ground-truth images (GTI) is critical, since DL relies heavily on large training datasets to achieve high accuracy.…”

Synthetization of high‐dose images using low‐dose CT scans

“…Deep learning (DL) has been explored in recent years for CT noise‐reduction as part of image reconstruction or post‐processing 19–28 . To further enhance the noise‐reduction capability, a dynamically selected image volume is utilized to produce a single image 29 .…”

Comparative Study of Traditional and Deep-Learning Denoising Approaches for Image-Based Petrophysical Characterization of Porous Media

“…The authors compared the effect of different loss functions and found that the least absolute deviation (L1) loss gave a better result in terms of peak signal to noise ratio (PSNR) and structural similarity index measure (SSIM) when compared to the least square errors (L2) loss, SSIM loss, and Visual Geometry Group (VGG) perceptual loss. Recently, Alsamadony et al (2021) acquired low-exposure and high-exposure CT scans on the same carbonate sample and compared denoising performance using a pre-trained very deep super resolution (VDSR) network against a shallow U-Net. They demonstrated that DL-based image processing can improve image quality; and that pre-trained VDSR network with fine-tuning tends to out-perform VDSR trained from scratch.…”