Reduction of Compton Background Noise for X-ray Fluorescence Computed Tomography with Deep Learning
For bench-top X-ray fluorescence computed tomography (XFCT), the X-ray tube source will bring extreme Compton background noise, resulting in a low signal-to-noise ratio and low contrast detection limit. In this paper, a noise2noise denoising algorithm based on the UNet deep learning network is propo...
Main Authors: | , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-02-01
|
Series: | Photonics |
Subjects: | |
Online Access: | https://www.mdpi.com/2304-6732/9/2/108 |
_version_ | 1797477110646833152 |
---|---|
author | Peng Feng Yan Luo Ruge Zhao Pan Huang Yonghui Li Peng He Bin Tang Xiansheng Zhao |
author_facet | Peng Feng Yan Luo Ruge Zhao Pan Huang Yonghui Li Peng He Bin Tang Xiansheng Zhao |
author_sort | Peng Feng |
collection | DOAJ |
description | For bench-top X-ray fluorescence computed tomography (XFCT), the X-ray tube source will bring extreme Compton background noise, resulting in a low signal-to-noise ratio and low contrast detection limit. In this paper, a noise2noise denoising algorithm based on the UNet deep learning network is proposed. The network can use noise image learning to convert the noise image into a clean image. Two sets of phantoms (high concentration Gd phantom and low concentration Bi phantom) are used for scanning to simulate the imaging process under different noise levels and generate the required data set. Additionally, the data set is generated by Geant4 simulation. In the training process, the L1 loss function is used for its good convergence. The image quality is evaluated according to CNR and pixel profile, which shows that our algorithm is better than BM3D, both visually and quantitatively. |
first_indexed | 2024-03-09T21:13:03Z |
format | Article |
id | doaj.art-790bf6cdd5eb4dc6a46308ff759e3336 |
institution | Directory Open Access Journal |
issn | 2304-6732 |
language | English |
last_indexed | 2024-03-09T21:13:03Z |
publishDate | 2022-02-01 |
publisher | MDPI AG |
record_format | Article |
series | Photonics |
spelling | doaj.art-790bf6cdd5eb4dc6a46308ff759e33362023-11-23T21:41:37ZengMDPI AGPhotonics2304-67322022-02-019210810.3390/photonics9020108Reduction of Compton Background Noise for X-ray Fluorescence Computed Tomography with Deep LearningPeng Feng0Yan Luo1Ruge Zhao2Pan Huang3Yonghui Li4Peng He5Bin Tang6Xiansheng Zhao7The Key Lab of Optoelectronic Technology and Systems, Ministry of Education, Chongqing University, Chongqing 400044, ChinaThe Key Lab of Optoelectronic Technology and Systems, Ministry of Education, Chongqing University, Chongqing 400044, ChinaThe Key Lab of Optoelectronic Technology and Systems, Ministry of Education, Chongqing University, Chongqing 400044, ChinaThe Key Lab of Optoelectronic Technology and Systems, Ministry of Education, Chongqing University, Chongqing 400044, ChinaThe Key Lab of Optoelectronic Technology and Systems, Ministry of Education, Chongqing University, Chongqing 400044, ChinaThe Key Lab of Optoelectronic Technology and Systems, Ministry of Education, Chongqing University, Chongqing 400044, ChinaThe Key Lab of Optoelectronic Technology and Systems, Ministry of Education, Chongqing University, Chongqing 400044, ChinaThe Key Lab of Optoelectronic Technology and Systems, Ministry of Education, Chongqing University, Chongqing 400044, ChinaFor bench-top X-ray fluorescence computed tomography (XFCT), the X-ray tube source will bring extreme Compton background noise, resulting in a low signal-to-noise ratio and low contrast detection limit. In this paper, a noise2noise denoising algorithm based on the UNet deep learning network is proposed. The network can use noise image learning to convert the noise image into a clean image. Two sets of phantoms (high concentration Gd phantom and low concentration Bi phantom) are used for scanning to simulate the imaging process under different noise levels and generate the required data set. Additionally, the data set is generated by Geant4 simulation. In the training process, the L1 loss function is used for its good convergence. The image quality is evaluated according to CNR and pixel profile, which shows that our algorithm is better than BM3D, both visually and quantitatively.https://www.mdpi.com/2304-6732/9/2/108XFCTCompton background noiseUNet networknoise2noise model |
spellingShingle | Peng Feng Yan Luo Ruge Zhao Pan Huang Yonghui Li Peng He Bin Tang Xiansheng Zhao Reduction of Compton Background Noise for X-ray Fluorescence Computed Tomography with Deep Learning Photonics XFCT Compton background noise UNet network noise2noise model |
title | Reduction of Compton Background Noise for X-ray Fluorescence Computed Tomography with Deep Learning |
title_full | Reduction of Compton Background Noise for X-ray Fluorescence Computed Tomography with Deep Learning |
title_fullStr | Reduction of Compton Background Noise for X-ray Fluorescence Computed Tomography with Deep Learning |
title_full_unstemmed | Reduction of Compton Background Noise for X-ray Fluorescence Computed Tomography with Deep Learning |
title_short | Reduction of Compton Background Noise for X-ray Fluorescence Computed Tomography with Deep Learning |
title_sort | reduction of compton background noise for x ray fluorescence computed tomography with deep learning |
topic | XFCT Compton background noise UNet network noise2noise model |
url | https://www.mdpi.com/2304-6732/9/2/108 |
work_keys_str_mv | AT pengfeng reductionofcomptonbackgroundnoiseforxrayfluorescencecomputedtomographywithdeeplearning AT yanluo reductionofcomptonbackgroundnoiseforxrayfluorescencecomputedtomographywithdeeplearning AT rugezhao reductionofcomptonbackgroundnoiseforxrayfluorescencecomputedtomographywithdeeplearning AT panhuang reductionofcomptonbackgroundnoiseforxrayfluorescencecomputedtomographywithdeeplearning AT yonghuili reductionofcomptonbackgroundnoiseforxrayfluorescencecomputedtomographywithdeeplearning AT penghe reductionofcomptonbackgroundnoiseforxrayfluorescencecomputedtomographywithdeeplearning AT bintang reductionofcomptonbackgroundnoiseforxrayfluorescencecomputedtomographywithdeeplearning AT xianshengzhao reductionofcomptonbackgroundnoiseforxrayfluorescencecomputedtomographywithdeeplearning |