Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Deep Convolutional Framelet Denosing for Low-Dose CT via Wavelet Residual Network
(Submitted on 31 Jul 2017 (v1), last revised 28 Mar 2018 (this version, v3))
Abstract: Model based iterative reconstruction (MBIR) algorithms for low-dose X-ray CT are computationally expensive. To address this problem, 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 were not fully recovered. To address this problem, here we propose a novel framelet-based denoising algorithm using wavelet residual network which synergistically combines the expressive power of deep learning and the performance guarantee from the framelet-based denoising algorithms. The new algorithms were inspired by the recent interpretation of the deep convolutional neural network (CNN) as a cascaded convolution framelet signal representation. Extensive experimental results confirm that the proposed networks have significantly improved performance and preserves the detail texture of the original images.
Submission history
From: Jong Chul Ye [view email][v1] Mon, 31 Jul 2017 16:17:31 GMT (2469kb,D)
[v2] Mon, 7 Aug 2017 16:10:04 GMT (2469kb,D)
[v3] Wed, 28 Mar 2018 07:46:15 GMT (4564kb,D)
Link back to: arXiv, form interface, contact.