STEDNet: Swin transformer‐based encoder–decoder network for noise reduction in low‐dose CT
Background Low‐dose computed tomography (LDCT) can reduce the dose of X‐ray radiation, making it increasingly significant for routine clinical diagnosis and treatment planning. However, the noise introduced by low‐dose X‐ray exposure degrades the quality of CT images, affecting the accuracy of clini...
Uloženo v:
| Vydáno v: | Medical physics (Lancaster) Ročník 50; číslo 7; s. 4443 - 4458 |
|---|---|
| Hlavní autoři: | , , , , , , , , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
United States
01.07.2023
|
| Témata: | |
| ISSN: | 0094-2405, 2473-4209, 2473-4209 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | Background
Low‐dose computed tomography (LDCT) can reduce the dose of X‐ray radiation, making it increasingly significant for routine clinical diagnosis and treatment planning. However, the noise introduced by low‐dose X‐ray exposure degrades the quality of CT images, affecting the accuracy of clinical diagnosis.
Purpose
The noises, artifacts, and high‐frequency components are similarly distributed in LDCT images. Transformer can capture global context information in an attentional manner to create distant dependencies on targets and extract more powerful features. In this paper, we reduce the impact of image errors on the ability to retain detailed information and improve the noise suppression performance by fully mining the distribution characteristics of image information.
Methods
This paper proposed an LDCT noise and artifact suppressing network based on Swin Transformer. The network includes a noise extraction sub‐network and a noise removal sub‐network. The noise extraction and removal capability are improved using a coarse extraction network of high‐frequency features based on full convolution. The noise removal sub‐network improves the network's ability to extract relevant image features by using a Swin Transformer with a shift window as an encoder–decoder and skip connections for global feature fusion. Also, the perceptual field is extended by extracting multi‐scale features of the images to recover the spatial resolution of the feature maps. The network uses a loss constraint with a combination of L1 and MS‐SSIM to improve and ensure the stability and denoising effect of the network.
Results
The denoising ability and clinical applicability of the methods were tested using clinical datasets. Compared with DnCNN, RED‐CNN, CBDNet and TSCN, the STEDNet method shows a better denoising effect on RMSE and PSNR. The STEDNet method effectively removes image noise and preserves the image structure to the maximum extent, making the reconstructed image closest to the NDCT image. The subjective and objective analysis of several sets of experiments shows that the method in this paper can effectively maintain the structure, edges, and textures of the denoised images while having good noise suppression performance. In the real data evaluation, the RMSE of this method is reduced by 18.82%, 15.15%, 2.25%, and 1.10% on average compared with DnCNN, RED‐CNN, CBDNet, and TSCNN, respectively. The average improvement of PSNR is 9.53%, 7.33%, 2.65%, and 3.69%, respectively.
Conclusions
This paper proposed a LDCT image denoising algorithm based on end‐to‐end training. The method in this paper can effectively improve the diagnostic performance of CT images by constraining the details of the images and restoring the LDCT image structure. The problem of increased noise and artifacts in CT images can be solved while maintaining the integrity of CT image tissue structure and pathological information. Compared with other algorithms, this method has better denoising effects both quantitatively and qualitatively. |
|---|---|
| AbstractList | Background
Low‐dose computed tomography (LDCT) can reduce the dose of X‐ray radiation, making it increasingly significant for routine clinical diagnosis and treatment planning. However, the noise introduced by low‐dose X‐ray exposure degrades the quality of CT images, affecting the accuracy of clinical diagnosis.
Purpose
The noises, artifacts, and high‐frequency components are similarly distributed in LDCT images. Transformer can capture global context information in an attentional manner to create distant dependencies on targets and extract more powerful features. In this paper, we reduce the impact of image errors on the ability to retain detailed information and improve the noise suppression performance by fully mining the distribution characteristics of image information.
Methods
This paper proposed an LDCT noise and artifact suppressing network based on Swin Transformer. The network includes a noise extraction sub‐network and a noise removal sub‐network. The noise extraction and removal capability are improved using a coarse extraction network of high‐frequency features based on full convolution. The noise removal sub‐network improves the network's ability to extract relevant image features by using a Swin Transformer with a shift window as an encoder–decoder and skip connections for global feature fusion. Also, the perceptual field is extended by extracting multi‐scale features of the images to recover the spatial resolution of the feature maps. The network uses a loss constraint with a combination of L1 and MS‐SSIM to improve and ensure the stability and denoising effect of the network.
Results
The denoising ability and clinical applicability of the methods were tested using clinical datasets. Compared with DnCNN, RED‐CNN, CBDNet and TSCN, the STEDNet method shows a better denoising effect on RMSE and PSNR. The STEDNet method effectively removes image noise and preserves the image structure to the maximum extent, making the reconstructed image closest to the NDCT image. The subjective and objective analysis of several sets of experiments shows that the method in this paper can effectively maintain the structure, edges, and textures of the denoised images while having good noise suppression performance. In the real data evaluation, the RMSE of this method is reduced by 18.82%, 15.15%, 2.25%, and 1.10% on average compared with DnCNN, RED‐CNN, CBDNet, and TSCNN, respectively. The average improvement of PSNR is 9.53%, 7.33%, 2.65%, and 3.69%, respectively.
Conclusions
This paper proposed a LDCT image denoising algorithm based on end‐to‐end training. The method in this paper can effectively improve the diagnostic performance of CT images by constraining the details of the images and restoring the LDCT image structure. The problem of increased noise and artifacts in CT images can be solved while maintaining the integrity of CT image tissue structure and pathological information. Compared with other algorithms, this method has better denoising effects both quantitatively and qualitatively. Low-dose computed tomography (LDCT) can reduce the dose of X-ray radiation, making it increasingly significant for routine clinical diagnosis and treatment planning. However, the noise introduced by low-dose X-ray exposure degrades the quality of CT images, affecting the accuracy of clinical diagnosis. Purpose The noises, artifacts, and high-frequency components are similarly distributed in LDCT images. Transformer can capture global context information in an attentional manner to create distant dependencies on targets and extract more powerful features. In this paper, we reduce the impact of image errors on the ability to retain detailed information and improve the noise suppression performance by fully mining the distribution characteristics of image information. This paper proposed an LDCT noise and artifact suppressing network based on Swin Transformer. The network includes a noise extraction sub-network and a noise removal sub-network. The noise extraction and removal capability are improved using a coarse extraction network of high-frequency features based on full convolution. The noise removal sub-network improves the network's ability to extract relevant image features by using a Swin Transformer with a shift window as an encoder-decoder and skip connections for global feature fusion. Also, the perceptual field is extended by extracting multi-scale features of the images to recover the spatial resolution of the feature maps. The network uses a loss constraint with a combination of L1 and MS-SSIM to improve and ensure the stability and denoising effect of the network. The denoising ability and clinical applicability of the methods were tested using clinical datasets. Compared with DnCNN, RED-CNN, CBDNet and TSCN, the STEDNet method shows a better denoising effect on RMSE and PSNR. The STEDNet method effectively removes image noise and preserves the image structure to the maximum extent, making the reconstructed image closest to the NDCT image. The subjective and objective analysis of several sets of experiments shows that the method in this paper can effectively maintain the structure, edges, and textures of the denoised images while having good noise suppression performance. In the real data evaluation, the RMSE of this method is reduced by 18.82%, 15.15%, 2.25%, and 1.10% on average compared with DnCNN, RED-CNN, CBDNet, and TSCNN, respectively. The average improvement of PSNR is 9.53%, 7.33%, 2.65%, and 3.69%, respectively. This paper proposed a LDCT image denoising algorithm based on end-to-end training. The method in this paper can effectively improve the diagnostic performance of CT images by constraining the details of the images and restoring the LDCT image structure. The problem of increased noise and artifacts in CT images can be solved while maintaining the integrity of CT image tissue structure and pathological information. Compared with other algorithms, this method has better denoising effects both quantitatively and qualitatively. Low-dose computed tomography (LDCT) can reduce the dose of X-ray radiation, making it increasingly significant for routine clinical diagnosis and treatment planning. However, the noise introduced by low-dose X-ray exposure degrades the quality of CT images, affecting the accuracy of clinical diagnosis. Purpose The noises, artifacts, and high-frequency components are similarly distributed in LDCT images. Transformer can capture global context information in an attentional manner to create distant dependencies on targets and extract more powerful features. In this paper, we reduce the impact of image errors on the ability to retain detailed information and improve the noise suppression performance by fully mining the distribution characteristics of image information.BACKGROUNDLow-dose computed tomography (LDCT) can reduce the dose of X-ray radiation, making it increasingly significant for routine clinical diagnosis and treatment planning. However, the noise introduced by low-dose X-ray exposure degrades the quality of CT images, affecting the accuracy of clinical diagnosis. Purpose The noises, artifacts, and high-frequency components are similarly distributed in LDCT images. Transformer can capture global context information in an attentional manner to create distant dependencies on targets and extract more powerful features. In this paper, we reduce the impact of image errors on the ability to retain detailed information and improve the noise suppression performance by fully mining the distribution characteristics of image information.This paper proposed an LDCT noise and artifact suppressing network based on Swin Transformer. The network includes a noise extraction sub-network and a noise removal sub-network. The noise extraction and removal capability are improved using a coarse extraction network of high-frequency features based on full convolution. The noise removal sub-network improves the network's ability to extract relevant image features by using a Swin Transformer with a shift window as an encoder-decoder and skip connections for global feature fusion. Also, the perceptual field is extended by extracting multi-scale features of the images to recover the spatial resolution of the feature maps. The network uses a loss constraint with a combination of L1 and MS-SSIM to improve and ensure the stability and denoising effect of the network.METHODSThis paper proposed an LDCT noise and artifact suppressing network based on Swin Transformer. The network includes a noise extraction sub-network and a noise removal sub-network. The noise extraction and removal capability are improved using a coarse extraction network of high-frequency features based on full convolution. The noise removal sub-network improves the network's ability to extract relevant image features by using a Swin Transformer with a shift window as an encoder-decoder and skip connections for global feature fusion. Also, the perceptual field is extended by extracting multi-scale features of the images to recover the spatial resolution of the feature maps. The network uses a loss constraint with a combination of L1 and MS-SSIM to improve and ensure the stability and denoising effect of the network.The denoising ability and clinical applicability of the methods were tested using clinical datasets. Compared with DnCNN, RED-CNN, CBDNet and TSCN, the STEDNet method shows a better denoising effect on RMSE and PSNR. The STEDNet method effectively removes image noise and preserves the image structure to the maximum extent, making the reconstructed image closest to the NDCT image. The subjective and objective analysis of several sets of experiments shows that the method in this paper can effectively maintain the structure, edges, and textures of the denoised images while having good noise suppression performance. In the real data evaluation, the RMSE of this method is reduced by 18.82%, 15.15%, 2.25%, and 1.10% on average compared with DnCNN, RED-CNN, CBDNet, and TSCNN, respectively. The average improvement of PSNR is 9.53%, 7.33%, 2.65%, and 3.69%, respectively.RESULTSThe denoising ability and clinical applicability of the methods were tested using clinical datasets. Compared with DnCNN, RED-CNN, CBDNet and TSCN, the STEDNet method shows a better denoising effect on RMSE and PSNR. The STEDNet method effectively removes image noise and preserves the image structure to the maximum extent, making the reconstructed image closest to the NDCT image. The subjective and objective analysis of several sets of experiments shows that the method in this paper can effectively maintain the structure, edges, and textures of the denoised images while having good noise suppression performance. In the real data evaluation, the RMSE of this method is reduced by 18.82%, 15.15%, 2.25%, and 1.10% on average compared with DnCNN, RED-CNN, CBDNet, and TSCNN, respectively. The average improvement of PSNR is 9.53%, 7.33%, 2.65%, and 3.69%, respectively.This paper proposed a LDCT image denoising algorithm based on end-to-end training. The method in this paper can effectively improve the diagnostic performance of CT images by constraining the details of the images and restoring the LDCT image structure. The problem of increased noise and artifacts in CT images can be solved while maintaining the integrity of CT image tissue structure and pathological information. Compared with other algorithms, this method has better denoising effects both quantitatively and qualitatively.CONCLUSIONSThis paper proposed a LDCT image denoising algorithm based on end-to-end training. The method in this paper can effectively improve the diagnostic performance of CT images by constraining the details of the images and restoring the LDCT image structure. The problem of increased noise and artifacts in CT images can be solved while maintaining the integrity of CT image tissue structure and pathological information. Compared with other algorithms, this method has better denoising effects both quantitatively and qualitatively. |
| Author | Zhu, Linlin Fu, Huijuan Yang, Shuangzhan Tan, Siyu Liu, Mengnan Xi, Xiaoqi Yan, Bin Han, Yu Li, Lei Liu, Chang |
| Author_xml | – sequence: 1 givenname: Linlin orcidid: 0000-0001-8932-9758 surname: Zhu fullname: Zhu, Linlin email: zhulinlingfs@163.com organization: PLA Strategic Support Force Information Engineering University – sequence: 2 givenname: Yu surname: Han fullname: Han, Yu organization: PLA Strategic Support Force Information Engineering University – sequence: 3 givenname: Xiaoqi surname: Xi fullname: Xi, Xiaoqi organization: PLA Strategic Support Force Information Engineering University – sequence: 4 givenname: Huijuan surname: Fu fullname: Fu, Huijuan organization: PLA Strategic Support Force Information Engineering University – sequence: 5 givenname: Siyu surname: Tan fullname: Tan, Siyu organization: PLA Strategic Support Force Information Engineering University – sequence: 6 givenname: Mengnan surname: Liu fullname: Liu, Mengnan organization: PLA Strategic Support Force Information Engineering University – sequence: 7 givenname: Shuangzhan surname: Yang fullname: Yang, Shuangzhan organization: PLA Strategic Support Force Information Engineering University – sequence: 8 givenname: Chang surname: Liu fullname: Liu, Chang organization: Zhengzhou University – sequence: 9 givenname: Lei surname: Li fullname: Li, Lei organization: PLA Strategic Support Force Information Engineering University – sequence: 10 givenname: Bin surname: Yan fullname: Yan, Bin email: ybspace@hotmail.com organization: PLA Strategic Support Force Information Engineering University |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/36708286$$D View this record in MEDLINE/PubMed |
| BookMark | eNp1kMtKxDAUhoMoOo6CTyBduul4mnZycSfjFbyB47qm6SlU22RMOgzufATBN_RJzMx4AVGyOEn4_p_Dt0lWjTVIyE4CgwSA7reTQcJoJldIj2Y8jTMKcpX0AGQW0wyGG2TT-wcAYOkQ1slGyjgIKliP3N-Oj4-usDuIbme1iTqnjK-sa9G9v7wWymMZodG2nL_fSlzcIoPdzLrHKICRsbXHyGE51V1tTRRKGjsL4dKG_9F4i6xVqvG4_Tn75O7keDw6iy-uT89HhxexTmkiY51QGPJClayqBOgUM0g51VBmVEgWjhJchv2BAyJTTMmiEinTQ6YLJlCmfbK37J04-zRF3-Vt7TU2jTJopz6nnEPGhaQioLuf6LRoscwnrm6Ve86_rPx0aWe9d1h9Iwnkc-F5O8kXwgM6-IXqulNzE0Fl3fwViJeBWd3g87_F-eXNkv8AvEaR-g |
| CitedBy_id | crossref_primary_10_1007_s10278_025_01390_0 crossref_primary_10_1088_1742_6596_2858_1_012038 crossref_primary_10_1109_JBHI_2024_3506785 crossref_primary_10_1007_s13534_024_00419_7 crossref_primary_10_1016_j_bspc_2024_107195 crossref_primary_10_1002_mp_17387 crossref_primary_10_1016_j_measurement_2024_115434 crossref_primary_10_1016_j_micron_2024_103615 crossref_primary_10_1007_s11760_023_02853_z crossref_primary_10_1007_s13534_024_00424_w crossref_primary_10_1016_j_nima_2025_170524 crossref_primary_10_1016_j_ceja_2025_100804 crossref_primary_10_1016_j_ejrad_2024_111355 crossref_primary_10_1007_s00117_023_01167_y crossref_primary_10_32604_cmc_2024_049623 crossref_primary_10_1088_1361_6560_addea6 crossref_primary_10_1007_s00034_024_02631_3 crossref_primary_10_1007_s10278_024_01254_z crossref_primary_10_1016_j_compmedimag_2025_102508 |
| Cites_doi | 10.1109/ACCESS.2020.3002534 10.1148/radiol.2311030191 10.1145/3065386 10.1162/neco.2006.18.7.1527 10.1002/mp.15176 10.1109/TMI.2020.2968472 10.1007/978-3-030-87231-1_6 10.1126/science.1127647 10.1109/TRPMS.2018.2860788 10.1088/1361-6560/ab325e 10.1109/CVPR.2019.00181 10.1109/TMI.2017.2715284 10.1148/radiology.175.3.2343122 10.1002/mp.14594 10.1109/TIP.2009.2017139 10.1007/978-3-319-24574-4_28 10.1002/mp.12344 10.1118/1.2230762 10.1117/1.JMI.5.3.036501 10.1007/s10462-020-09861-2 10.1109/ACCESS.2020.3006512 10.1088/1361-6560/acc000 10.1007/s10278-018-0056-0 10.1109/TIP.2017.2662206 10.1109/TIM.2019.2925881 10.1109/ICCV48922.2021.00986 10.1109/TMI.2004.838324 10.1109/TCSVT.2020.2988895 10.1109/TCSVT.2020.2982174 10.1007/s10278-019-00274-4 10.1109/TMI.2018.2827462 |
| ContentType | Journal Article |
| Copyright | 2023 American Association of Physicists in Medicine. |
| Copyright_xml | – notice: 2023 American Association of Physicists in Medicine. |
| DBID | AAYXX CITATION CGR CUY CVF ECM EIF NPM 7X8 |
| DOI | 10.1002/mp.16249 |
| DatabaseName | CrossRef Medline MEDLINE MEDLINE (Ovid) MEDLINE MEDLINE PubMed MEDLINE - Academic |
| DatabaseTitle | CrossRef MEDLINE Medline Complete MEDLINE with Full Text PubMed MEDLINE (Ovid) MEDLINE - Academic |
| DatabaseTitleList | MEDLINE MEDLINE - Academic |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | fulltext_linktorsrc |
| Discipline | Medicine Physics |
| EISSN | 2473-4209 |
| EndPage | 4458 |
| ExternalDocumentID | 36708286 10_1002_mp_16249 MP16249 |
| Genre | article Journal Article |
| GrantInformation_xml | – fundername: National Key Research and Development Program of China funderid: 2020YFC1522002 – fundername: National Natural Science Foundation of China funderid: 62201618 – fundername: National Natural Science Foundation of China grantid: 62201618 – fundername: National Key Research and Development Program of China grantid: 2020YFC1522002 |
| GroupedDBID | --- --Z -DZ .GJ 0R~ 1OB 1OC 29M 2WC 33P 36B 3O- 4.4 53G 5GY 5RE 5VS AAHHS AAHQN AAIPD AAMNL AANLZ AAQQT AASGY AAXRX AAYCA AAZKR ABCUV ABDPE ABEFU ABFTF ABJNI ABLJU ABQWH ABTAH ABXGK ACAHQ ACBEA ACCFJ ACCZN ACGFO ACGFS ACGOF ACPOU ACXBN ACXQS ADBBV ADBTR ADKYN ADOZA ADXAS ADZMN AEEZP AEGXH AEIGN AENEX AEQDE AEUYR AFBPY AFFPM AFWVQ AHBTC AIACR AIAGR AITYG AIURR AIWBW AJBDE ALMA_UNASSIGNED_HOLDINGS ALUQN ALVPJ AMYDB ASPBG BFHJK C45 CS3 DCZOG DRFUL DRMAN DRSTM DU5 EBD EBS EJD EMB EMOBN F5P HDBZQ HGLYW I-F KBYEO LATKE LEEKS LOXES LUTES LYRES MEWTI O9- OVD P2P P2W PALCI PHY RJQFR RNS ROL SAMSI SUPJJ SV3 TEORI TN5 TWZ USG WOHZO WXSBR XJT ZGI ZVN ZXP ZY4 ZZTAW AAMMB AAYXX ABUFD ADMLS AEFGJ AEYWJ AGHNM AGXDD AGYGG AIDQK AIDYY AIQQE CITATION LH4 CGR CUY CVF ECM EIF NPM 7X8 |
| ID | FETCH-LOGICAL-c3219-c12057bad6ff80c3e40372c0d42896969a879006070ee6a6a9bf836c56cb68e93 |
| IEDL.DBID | DRFUL |
| ISICitedReferencesCount | 29 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000928740400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 0094-2405 2473-4209 |
| IngestDate | Fri Sep 05 11:46:31 EDT 2025 Mon Jul 21 06:00:39 EDT 2025 Sat Nov 29 06:02:53 EST 2025 Tue Nov 18 22:37:41 EST 2025 Wed Jan 22 16:21:43 EST 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 7 |
| Keywords | low-dose CT Swin transformer deep learning encoder-decoder |
| Language | English |
| License | 2023 American Association of Physicists in Medicine. |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c3219-c12057bad6ff80c3e40372c0d42896969a879006070ee6a6a9bf836c56cb68e93 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| ORCID | 0000-0001-8932-9758 |
| PMID | 36708286 |
| PQID | 2770478928 |
| PQPubID | 23479 |
| PageCount | 16 |
| ParticipantIDs | proquest_miscellaneous_2770478928 pubmed_primary_36708286 crossref_primary_10_1002_mp_16249 crossref_citationtrail_10_1002_mp_16249 wiley_primary_10_1002_mp_16249_MP16249 |
| PublicationCentury | 2000 |
| PublicationDate | July 2023 2023-07-00 2023-Jul 20230701 |
| PublicationDateYYYYMMDD | 2023-07-01 |
| PublicationDate_xml | – month: 07 year: 2023 text: July 2023 |
| PublicationDecade | 2020 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | Medical physics (Lancaster) |
| PublicationTitleAlternate | Med Phys |
| PublicationYear | 2023 |
| References | 2021; 48 2019; 3 2017; 60 2017; 26 2006; 33 2017; 44 2020; 39 2006; 18 2020; 33 2006; 313 2005; 24 2020; 8 2015; 9351 2004; 231 2021; 54 2021; 31 2018; 5 2019; 64 2017; 36 2022 2021 2019 2020; 69 2017 2014 2018; 31 1990; 175 2018; 37 2009; 18 e_1_2_7_6_1 e_1_2_7_5_1 e_1_2_7_4_1 e_1_2_7_3_1 e_1_2_7_9_1 e_1_2_7_8_1 e_1_2_7_7_1 e_1_2_7_19_1 e_1_2_7_18_1 e_1_2_7_17_1 e_1_2_7_16_1 e_1_2_7_2_1 e_1_2_7_15_1 e_1_2_7_14_1 e_1_2_7_13_1 e_1_2_7_12_1 e_1_2_7_11_1 e_1_2_7_10_1 e_1_2_7_26_1 e_1_2_7_27_1 e_1_2_7_28_1 e_1_2_7_29_1 e_1_2_7_30_1 e_1_2_7_25_1 e_1_2_7_31_1 e_1_2_7_24_1 e_1_2_7_32_1 e_1_2_7_23_1 e_1_2_7_33_1 e_1_2_7_22_1 e_1_2_7_34_1 e_1_2_7_21_1 e_1_2_7_35_1 e_1_2_7_20_1 e_1_2_7_36_1 e_1_2_7_37_1 |
| References_xml | – volume: 64 year: 2019 article-title: Gradient regularized convolutional neural networks for low‐dose CT image enhancement publication-title: Phys Med Biol – volume: 31 start-page: 655 year: 2018 end-page: 669 article-title: Sharpness‐aware low‐dose CT denoising using conditional generative adversarial network publication-title: J Digit Imaging – volume: 44 start-page: e360 year: 2017 end-page: e375 article-title: A deep convolutional neural network using directional wavelets for low‐dose X‐ray CT reconstruction publication-title: Med Phys – volume: 69 start-page: 2707 year: 2020 end-page: 2721 article-title: Two‐stage convolutional neural network for medical noise removal via image decomposition publication-title: IEEE Trans Instrum Meas – volume: 3 start-page: 137 year: 2019 end-page: 152 article-title: Convolutional neural network‐based robust denoising of low‐dose computed tomography perfusion maps publication-title: IEEE Trans Radiat Plasma Med Sci – volume: 175 start-page: 729 year: 1990 end-page: 731 article-title: Low‐dose CT of the lungs: preliminary observations publication-title: Radiology – volume: 8 start-page: 112 078 year: 2020 end-page: 112 091 article-title: Image restoration for low‐dose ct via transfer learning and residual network publication-title: IEEE Access – volume: 48 start-page: 902 year: 2021 end-page: 911 article-title: Low‐dose CT image and projection dataset publication-title: Med Phys – volume: 24 start-page: 105 year: 2005 end-page: 111 article-title: Reduction of noise‐induced streak artifacts in X‐ray computed tomography through spline‐based penalized‐likelihood sinogram smoothing publication-title: IEEE Trans Med Imaging – volume: 37 start-page: 1348 year: 2018 end-page: 1357 article-title: Low‐dose ct image denoising using a generative adversarial network with wasserstein distance and perceptual loss publication-title: IEEE Trans Med Imaging – year: 2021 – volume: 18 start-page: 1228 year: 2009 end-page: 1238 article-title: Electronic noise modeling in statistical iterative reconstruction publication-title: IEEE Trans Image Process – start-page: 1712 year: 2019 end-page: 1722 article-title: Toward convolutional blind denoising of real photographs – volume: 54 start-page: 215 year: 2021 end-page: 251 article-title: Deep learning for biomedical image reconstruction: a survey publication-title: Artif Intell Rev – volume: 39 start-page: 2289 year: 2020 end-page: 2301 article-title: SACNN: Self‐attention convolutional neural network for low‐dose CT denoising with self‐supervised perceptual loss network publication-title: IEEE Trans Med Imaging – volume: 36 start-page: 2524 year: 2017 end-page: 2535 article-title: Low‐dose CT with a residual encoder‐decoder convolutional neural network publication-title: IEEE Trans Med Imaging – volume: 33 start-page: 504 year: 2020 end-page: 515 article-title: Deep learning for low‐dose CT denoising using perceptual loss and edge detection layer publication-title: J Digit Imaging – volume: 8 start-page: 133 470 year: 2020 end-page: 133 487 article-title: Single low‐dose CT image denoising using a generative adversarial network with modified U‐Net Generator and multi‐level discriminator publication-title: IEEE Access – start-page: 55 year: 2021 end-page: 64 – volume: 31 start-page: 467 year: 2021 end-page: 479 article-title: Dual‐stream multi‐path recursive residual network for JPEG image compression artifacts reduction publication-title: IEEE Trans Circuits Syst Video Technol – volume: 48 start-page: 5712 year: 2021 end-page: 5726 article-title: Deep learning enabled ultra‐fast‐pitch acquisition in clinical X‐ray computed tomography publication-title: Med Phys – volume: 231 start-page: 169 year: 2004 end-page: 174 article-title: Low‐kilovoltage multi‐detector row chest CT in adults: feasibility and effect on image quality and iodine dose publication-title: Radiology – volume: 31 start-page: 512 year: 2021 end-page: 522 article-title: Multi‐grained attention networks for single image super‐resolution publication-title: IEEE Trans Circuits Syst Video Technol – volume: 26 start-page: 3142 year: 2017 end-page: 3155 article-title: Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising publication-title: IEEE Trans Image Process – year: 2022 – volume: 33 start-page: 3290 year: 2006 end-page: 3303 article-title: Properties of preprocessed sinogram data in x‐ray computed tomography: properties of preprocessed sinogram data in x‐ray CT publication-title: Med Phys – volume: 313 start-page: 504 year: 2006 end-page: 507 article-title: Reducing the dimensionality of data with neural networks publication-title: Science – start-page: 2672 year: 2014 end-page: 2680 – volume: 60 start-page: 84 year: 2017 end-page: 90 article-title: ImageNet classification with deep convolutional neural networks publication-title: Commun ACM – year: 2017 – year: 2019 – volume: 9351 start-page: 234 year: 2015 end-page: 241 – volume: 18 start-page: 1527 year: 2006 end-page: 1554 article-title: A fast learning algorithm for deep belief nets publication-title: Neural Comput – volume: 5 start-page: 12 year: 2018 article-title: DeepLesion: Automated mining of large‐scale lesion annotations and universal lesion detection with deep learning publication-title: J Med Imaging – ident: e_1_2_7_17_1 doi: 10.1109/ACCESS.2020.3002534 – ident: e_1_2_7_3_1 doi: 10.1148/radiol.2311030191 – ident: e_1_2_7_5_1 doi: 10.1145/3065386 – ident: e_1_2_7_11_1 doi: 10.1162/neco.2006.18.7.1527 – ident: e_1_2_7_25_1 doi: 10.1002/mp.15176 – ident: e_1_2_7_29_1 – ident: e_1_2_7_23_1 doi: 10.1109/TMI.2020.2968472 – ident: e_1_2_7_27_1 doi: 10.1007/978-3-030-87231-1_6 – ident: e_1_2_7_10_1 doi: 10.1126/science.1127647 – ident: e_1_2_7_14_1 doi: 10.1109/TRPMS.2018.2860788 – ident: e_1_2_7_16_1 doi: 10.1088/1361-6560/ab325e – ident: e_1_2_7_37_1 doi: 10.1109/CVPR.2019.00181 – ident: e_1_2_7_12_1 doi: 10.1109/TMI.2017.2715284 – ident: e_1_2_7_2_1 doi: 10.1148/radiology.175.3.2343122 – ident: e_1_2_7_35_1 doi: 10.1002/mp.14594 – ident: e_1_2_7_33_1 doi: 10.1109/TIP.2009.2017139 – ident: e_1_2_7_6_1 doi: 10.1007/978-3-319-24574-4_28 – ident: e_1_2_7_7_1 – ident: e_1_2_7_13_1 doi: 10.1002/mp.12344 – ident: e_1_2_7_31_1 doi: 10.1118/1.2230762 – ident: e_1_2_7_34_1 doi: 10.1117/1.JMI.5.3.036501 – ident: e_1_2_7_4_1 doi: 10.1007/s10462-020-09861-2 – ident: e_1_2_7_22_1 doi: 10.1109/ACCESS.2020.3006512 – ident: e_1_2_7_28_1 doi: 10.1088/1361-6560/acc000 – ident: e_1_2_7_21_1 doi: 10.1007/s10278-018-0056-0 – ident: e_1_2_7_24_1 – ident: e_1_2_7_36_1 doi: 10.1109/TIP.2017.2662206 – ident: e_1_2_7_18_1 doi: 10.1109/TIM.2019.2925881 – ident: e_1_2_7_19_1 – ident: e_1_2_7_30_1 doi: 10.1109/ICCV48922.2021.00986 – ident: e_1_2_7_32_1 doi: 10.1109/TMI.2004.838324 – ident: e_1_2_7_9_1 doi: 10.1109/TCSVT.2020.2988895 – ident: e_1_2_7_26_1 – ident: e_1_2_7_8_1 doi: 10.1109/TCSVT.2020.2982174 – ident: e_1_2_7_15_1 doi: 10.1007/s10278-019-00274-4 – ident: e_1_2_7_20_1 doi: 10.1109/TMI.2018.2827462 |
| SSID | ssj0006350 |
| Score | 2.5368521 |
| Snippet | Background
Low‐dose computed tomography (LDCT) can reduce the dose of X‐ray radiation, making it increasingly significant for routine clinical diagnosis and... Low-dose computed tomography (LDCT) can reduce the dose of X-ray radiation, making it increasingly significant for routine clinical diagnosis and treatment... |
| SourceID | proquest pubmed crossref wiley |
| SourceType | Aggregation Database Index Database Enrichment Source Publisher |
| StartPage | 4443 |
| SubjectTerms | Algorithms deep learning encoder–decoder Image Processing, Computer-Assisted - methods low‐dose CT Radiation Dosage Signal-To-Noise Ratio Swin transformer Tomography, X-Ray Computed - methods |
| Title | STEDNet: Swin transformer‐based encoder–decoder network for noise reduction in low‐dose CT |
| URI | https://onlinelibrary.wiley.com/doi/abs/10.1002%2Fmp.16249 https://www.ncbi.nlm.nih.gov/pubmed/36708286 https://www.proquest.com/docview/2770478928 |
| Volume | 50 |
| WOSCitedRecordID | wos000928740400001&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | 1 |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| journalDatabaseRights | – providerCode: PRVWIB databaseName: Wiley Online Library Full Collection 2020 customDbUrl: eissn: 2473-4209 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0006350 issn: 0094-2405 databaseCode: DRFUL dateStart: 19970101 isFulltext: true titleUrlDefault: https://onlinelibrary.wiley.com providerName: Wiley-Blackwell |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1La9wwEB7yaEsvSZo-smkaVCjtyYnX8uqRW0iy9NAsIQ_YmytLY1hI7GW9Sa79CYX-w_6Sjix7S2gKhZ78YCQbz2ge1sw3AB_UoDBG8zQS0hYUoEge5X2jIi5M6nJhyYcNzSbkaKTGY33WZlX6WpiAD7H44eZXRqOv_QI3eb3_GzT0ZrrXFxQ8LMOqr6miwGv1-Hx49WWhh8mUhgIUnfo9hEEHPRsn-93Yh8boDw_zocPaWJzh-v-86wastX4mOwyC8QKWsNyEZ6ftTvomPG1SP239Er5ekPIa4fyAXdxPSjbvPFmc_fz23Rs5xzzYpfPXPxw2Z6wM2eOMCFlZTWpkM48B67nMaJLr6p4Gu4ruH12-gqvhyeXR56jtuxBZTgossv2EvLjcOFEUKrYc05jLxMaOQhUPpqONktoDucgYURhhdF4oTnwVNhcKNX8NK2VV4hYwdCQHBWLKUacmH1Bw1ecyljJ2MWqT9OBTx4DMtqDkvjfGdRbglJPsZpo1n64H7xeU0wDE8RhNx8OMVonf-jAlVrd1ltATU6l0onrwJjB3MYuHsPPF9D342PDwr9Nnp2fNcftfCd_Cc9-dPmT37sDKfHaL7-CJvZtP6tkuLMux2m2l9hflF-8X |
| linkProvider | Wiley-Blackwell |
| linkToHtml | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1Jb9NAFH4qKZReWAqlYR0kRE-mjseZBU6oEBWRRBVNpd7MeOZZitTaUZK2V34CUv8hv4Q3HjuoAiQkTl70Zsby22f5HsAr1S-M0TyNhLQFJSiSR3nPqIgLk7pcWIphQ7EJOR6rkxN9uAbv2rMwAR9iNeHmNaO2117B_YT03i_U0LPZm56g7OEGrKeCS9WB9Q9fBsfDlSEmXxpOoOjULyL0W-zZONlr2173Rr-FmNcj1trlDO7-18fegztNpMneB9G4D2tYbsHGqFlL34Jb9eZPu3gAX4_IfI1x-ZYdXU5LtmxjWZz_-PbduznHPNyl889XDus7Vob944wIWVlNF8jmHgXW85lRJ6fVJTV2Fb3fnzyE48HHyf5B1FReiCwnExbZXkJxXG6cKAoVW45pzGViY0fJiofT0UZJ7aFcZIwojDA6LxQnzgqbC4Wab0OnrErcAYaOJKFATDnq1OR9Sq96XMZSxi5GbZIu7LYcyGwDS-6rY5xmAVA5yc5mWf3ruvByRTkLUBx_ommZmJGe-MUPU2J1vsgSGjGVSieqC48Cd1e9eBA7f5y-C69rJv61-2x0WF8f_yvhC7h9MBkNs-Gn8ecnsOlr1Ye9vk-hs5yf4zO4aS-W08X8eSO8PwGRWvIf |
| linkToPdf | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LT9wwEB7RpUVcgNIHC5S6UtWeUrKx4wc9VdAVVWG1KiBxC449kVaCZLW7lGt_QqX-Q34JdpxshdpKlTjlobETZcbziGe-AXgr00JrRVnEhSlcgCJolPe0jCjXzObcOB82NJsQg4E8P1fDBfjY1sIEfIj5Dze_Mmp97Rc4jm2x-xs19Gr8ocdd9PAIFlmqUtaBxYNv_bOjuSJ2tjRUoCjmNxHSFns2Tnbbsfet0R8u5n2PtTY5_dUHvewarDSeJvkUROMpLGC5DkvHzV76Ojypkz_N9BlcnDj1NcDZHjm5GZVk1vqyOLn98dObOUs83KX1178s1mekDPnjxBGSshpNkUw8CqznM3GTXFY3brCt3P390-dw1v98un8YNZ0XIkOdCotML3F-XK4tLwoZG4ospiIxsXXBiofTUVoK5aFcRIzINdcqLyR1nOUm5xIVfQGdsipxAwhaJwkFIqOomM5TF171qIiFiG2MSiddeN9yIDMNLLnvjnGZBUDlJLsaZ_Wn68KbOeU4QHH8jaZlYubWid_80CVW19MscU9kQqpEduFl4O58Fg9i58vpu_CuZuI_p8-Oh_Vx838JX8PS8KCfHX0ZfN2CZd-qPqT6bkNnNrnGV_DYfJ-NppOdRnbvADF28Zo |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=STEDNet%3A+Swin+transformer%E2%80%90based+encoder%E2%80%93decoder+network+for+noise+reduction+in+low%E2%80%90dose+CT&rft.jtitle=Medical+physics+%28Lancaster%29&rft.au=Zhu%2C+Linlin&rft.au=Han%2C+Yu&rft.au=Xi%2C+Xiaoqi&rft.au=Fu%2C+Huijuan&rft.date=2023-07-01&rft.issn=0094-2405&rft.eissn=2473-4209&rft.volume=50&rft.issue=7&rft.spage=4443&rft.epage=4458&rft_id=info:doi/10.1002%2Fmp.16249&rft.externalDBID=10.1002%252Fmp.16249&rft.externalDocID=MP16249 |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0094-2405&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0094-2405&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0094-2405&client=summon |