Wavelet Domain Residual Network (WavResNet) for Low-Dose X-ray CT Reconstruction

TL;DR

This paper introduces WavResNet, a wavelet domain residual network that enhances low-dose X-ray CT reconstruction by effectively denoising wavelet coefficients, leading to improved detail preservation over previous CNN methods.

Contribution

It proposes a novel wavelet domain residual learning approach that directly estimates and removes noise from wavelet coefficients, improving texture recovery in low-dose CT images.

Findings

Significantly improved image detail preservation.

Enhanced noise reduction in wavelet domain.

Outperforms previous CNN-based methods.

Abstract

Model based iterative reconstruction (MBIR) algorithms for low-dose X-ray CT are computationally complex because of the repeated use of the forward and backward projection. Inspired by this success of deep learning in computer vision applications, we recently proposed a deep convolutional neural network (CNN) for low-dose X-ray CT and won the second place in 2016 AAPM Low-Dose CT Grand Challenge. However, some of the texture are not fully recovered, which was unfamiliar to some radiologists. To cope with this problem, here we propose a direct residual learning approach on directional wavelet domain to solve this problem and to improve the performance against previous work. In particular, the new network estimates the noise of each input wavelet transform, and then the de-noised wavelet coefficients are obtained by subtracting the noise from the input wavelet transform bands. The experimental results confirm that the proposed network has significantly improved performance, preserving the detail texture of the original images.