SART-Res-UNet: Fan Beam CT Image Reconstruction from Limited Projections using attention-enabled residual U-Net

Harika Jinka, Jyothsna Shaji, Sangeeth John, Sreeraj R Menon, Amalu Pradeep, Jayaraj P B, Pournami P N, Niyas Puzhakkal
Proceedings of the 15th Asian Conference on Machine Learning, PMLR 222:598-613, 2024.

Abstract

CT scans significantly improve analytical competencies but uses X-Rays which will produce ionizing radiation that bring higher radiation to the living tissues. Thus, optimization of CT radiation dose has a significant concern to lower the health risks. Many manufacturers have done a greater contribution by developing technologies to reduce dosage by maintaining image quality by adding noise reduction filters, automatic exposure control, and using many iterative reconstruction algorithms. Image reconstruction algorithms play a vital role in maintaining or improving image quality in reduced-dose CT. The present research work combines the state-of-the-art reconstruction technique Simultaneous Algebraic Reconstruction Technique (SART) with a Residual U-Net network to generate images from limited number of sinograms. The proposed model is trained using sinograms corresponding head and neck and head CT images of 10 patients. The proposed model predicted superior diagnostic quality images with max PSNR of 70.23 and Structural Similarity Index Measure (SSIM) of 0.99. Thus the proposed model, SART-Res-Unet, ensures a very low radiation exposure to a patient during the repeated CT imaging sequence, which is an inevitable part of radiotherapy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v222-jinka24a, title = {{SART-Res-UNet}: {F}an Beam {CT} Image Reconstruction from Limited Projections using attention-enabled residual {U-Net}}, author = {Jinka, Harika and Shaji, Jyothsna and John, Sangeeth and Menon, Sreeraj R and Pradeep, Amalu and P B, Jayaraj and P N, Pournami and Puzhakkal, Niyas}, booktitle = {Proceedings of the 15th Asian Conference on Machine Learning}, pages = {598--613}, year = {2024}, editor = {Yanıkoğlu, Berrin and Buntine, Wray}, volume = {222}, series = {Proceedings of Machine Learning Research}, month = {11--14 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v222/jinka24a/jinka24a.pdf}, url = {https://proceedings.mlr.press/v222/jinka24a.html}, abstract = {CT scans significantly improve analytical competencies but uses X-Rays which will produce ionizing radiation that bring higher radiation to the living tissues. Thus, optimization of CT radiation dose has a significant concern to lower the health risks. Many manufacturers have done a greater contribution by developing technologies to reduce dosage by maintaining image quality by adding noise reduction filters, automatic exposure control, and using many iterative reconstruction algorithms. Image reconstruction algorithms play a vital role in maintaining or improving image quality in reduced-dose CT. The present research work combines the state-of-the-art reconstruction technique Simultaneous Algebraic Reconstruction Technique (SART) with a Residual U-Net network to generate images from limited number of sinograms. The proposed model is trained using sinograms corresponding head and neck and head CT images of 10 patients. The proposed model predicted superior diagnostic quality images with max PSNR of 70.23 and Structural Similarity Index Measure (SSIM) of 0.99. Thus the proposed model, SART-Res-Unet, ensures a very low radiation exposure to a patient during the repeated CT imaging sequence, which is an inevitable part of radiotherapy.} }
Endnote
%0 Conference Paper %T SART-Res-UNet: Fan Beam CT Image Reconstruction from Limited Projections using attention-enabled residual U-Net %A Harika Jinka %A Jyothsna Shaji %A Sangeeth John %A Sreeraj R Menon %A Amalu Pradeep %A Jayaraj P B %A Pournami P N %A Niyas Puzhakkal %B Proceedings of the 15th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Berrin Yanıkoğlu %E Wray Buntine %F pmlr-v222-jinka24a %I PMLR %P 598--613 %U https://proceedings.mlr.press/v222/jinka24a.html %V 222 %X CT scans significantly improve analytical competencies but uses X-Rays which will produce ionizing radiation that bring higher radiation to the living tissues. Thus, optimization of CT radiation dose has a significant concern to lower the health risks. Many manufacturers have done a greater contribution by developing technologies to reduce dosage by maintaining image quality by adding noise reduction filters, automatic exposure control, and using many iterative reconstruction algorithms. Image reconstruction algorithms play a vital role in maintaining or improving image quality in reduced-dose CT. The present research work combines the state-of-the-art reconstruction technique Simultaneous Algebraic Reconstruction Technique (SART) with a Residual U-Net network to generate images from limited number of sinograms. The proposed model is trained using sinograms corresponding head and neck and head CT images of 10 patients. The proposed model predicted superior diagnostic quality images with max PSNR of 70.23 and Structural Similarity Index Measure (SSIM) of 0.99. Thus the proposed model, SART-Res-Unet, ensures a very low radiation exposure to a patient during the repeated CT imaging sequence, which is an inevitable part of radiotherapy.
APA
Jinka, H., Shaji, J., John, S., Menon, S.R., Pradeep, A., P B, J., P N, P. & Puzhakkal, N.. (2024). SART-Res-UNet: Fan Beam CT Image Reconstruction from Limited Projections using attention-enabled residual U-Net. Proceedings of the 15th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 222:598-613 Available from https://proceedings.mlr.press/v222/jinka24a.html.

Related Material