In order to reduce the potential radiation risk, low-dose CT has gained increased attention in medical imaging community. Currently, patients go through multiple x-ray CT scans during image-guided radiation therapy, which elevates the potential risk for tissue damage and radiation-induced cancer [1, 2]. However, simply lowering the radiation dose will significantly degrade the image quality. Therefore, there is increasing demand for fast image reconstruction algorithms that can produce higher quality images in clinically relevant time. In this chapter, we propose a new noise reduction method for low-dose CT via deep learning without accessing original projection data. A deep convolutional neural network is used to map low-dose CT images towards its corresponding normal-dose counterparts using recently proposed residual learning method . Qualitative results demonstrate a great potential of the proposed method on artifact reduction and structure preservation. In terms of the quantitative metrics, the proposed method has showed a substantial improvement on PSNR, RMSE and SSIM than the competing state-of-art methods like Block Matching 3D (BM3D)  and Weighted Nuclear Norm Minimization (WNNM) . Furthermore, the speed of our method is significantly faster than the iterative and linear reconstruction methods discussed in previous chapters.
7.1.1 Deep Neural Networks for x-ray Image Denoising
Most clinical x-ray CT scanners currently being used employ some version of analytical reconstruction algorithms like FBP or FDK. However, in low dose x-ray CT, the linear reconstruction algorithms introduce severe artifacts typically due to beam hardening, photon starvation, scatter and other causes which reduces the diagnostic reliability. Therefore, high quality diagnostically relevant low dose x-ray CT reconstruction is a topic of major research effort. In previous chapters, we have observed that model based image reconstruction problems perform reliably well but they are still computationally expensive even with the introduction of multiple GPUs in parallel. As a result, we have explored the possibility of leveraging tremendous potential of artificial intelligence especially deep convolutional neural networks to perform x-ray CT image denoising.
Get Help With Your Dissertation
UKDiss.com provide professional assistance for university students of all levels. Our dissertation writing service…
You can read the article in its entirety, on the official website of https://ukdiss.com/examples/deep-convolutional-neural-network-based-denoising.php
Thanks a lot … I hope your thoughts, your geniuses can have as much resonance as possible.