CT-X-Ray Registration Via Spatial-Projective Dual Transformer Network Fused with Target Detection

Zheng Zhang; Danni Ai; Haixiao Geng; Jian Yang

doi:10.1145/3571532.3571546

CT-X-Ray Registration Via Spatial-Projective Dual Transformer Network Fused with Target Detection

Zheng Zhang, Danni Ai^*, Haixiao Geng, Jian Yang

^*Corresponding author for this work

School of Optics and Photonics

Beijing Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Registration of CT-X-rays is crucial in high-precision orthopedic surgery. In this study, a deep learning network integrating convolution and transformer modules is proposed as a model for measuring image similarity for the registration of CT-X-rays. By training the network model to approximate the geodesic distance of Riemann space, the model has the property of convex function, to avoid falling into a local optimum. To further reduce the translation error of registration, this study introduces a spine detection network based on Yolov5, detects the spine of the target image and the image to be registered, obtains the spine position information and readjusts the translation component of the pose. The method used in this study has been tested, and the translation error and rotation error are lower than 3.05 mm and 1.96°, respectively.

Original language	English
Title of host publication	ICBBS 2022 - 2022 11th International Conference on Bioinformatics and Biomedical Science
Publisher	Association for Computing Machinery
Pages	93-98
Number of pages	6
ISBN (Electronic)	9781450396929
DOIs	https://doi.org/10.1145/3571532.3571546
Publication status	Published - 28 Oct 2022
Event	11th International Conference on Bioinformatics and Biomedical Science, ICBBS 2022 - Nanning, China Duration: 28 Oct 2022 → 30 Oct 2022

Publication series

Name	ACM International Conference Proceeding Series

Conference

Conference	11th International Conference on Bioinformatics and Biomedical Science, ICBBS 2022
Country/Territory	China
City	Nanning
Period	28/10/22 → 30/10/22

Keywords

CT-Xray registration
Deep learning
Spine detection
Vision transformer

Access to Document

10.1145/3571532.3571546

Cite this

Zhang, Z., Ai, D., Geng, H., & Yang, J. (2022). CT-X-Ray Registration Via Spatial-Projective Dual Transformer Network Fused with Target Detection. In ICBBS 2022 - 2022 11th International Conference on Bioinformatics and Biomedical Science (pp. 93-98). (ACM International Conference Proceeding Series). Association for Computing Machinery. https://doi.org/10.1145/3571532.3571546

@inproceedings{23e5a23c135341f3be34a8cded0da8b7,

title = "CT-X-Ray Registration Via Spatial-Projective Dual Transformer Network Fused with Target Detection",

abstract = "Registration of CT-X-rays is crucial in high-precision orthopedic surgery. In this study, a deep learning network integrating convolution and transformer modules is proposed as a model for measuring image similarity for the registration of CT-X-rays. By training the network model to approximate the geodesic distance of Riemann space, the model has the property of convex function, to avoid falling into a local optimum. To further reduce the translation error of registration, this study introduces a spine detection network based on Yolov5, detects the spine of the target image and the image to be registered, obtains the spine position information and readjusts the translation component of the pose. The method used in this study has been tested, and the translation error and rotation error are lower than 3.05 mm and 1.96°, respectively.",

keywords = "CT-Xray registration, Deep learning, Spine detection, Vision transformer",

author = "Zheng Zhang and Danni Ai and Haixiao Geng and Jian Yang",

note = "Publisher Copyright: {\textcopyright} 2022 ACM.; 11th International Conference on Bioinformatics and Biomedical Science, ICBBS 2022 ; Conference date: 28-10-2022 Through 30-10-2022",

year = "2022",

month = oct,

day = "28",

doi = "10.1145/3571532.3571546",

language = "English",

series = "ACM International Conference Proceeding Series",

publisher = "Association for Computing Machinery",

pages = "93--98",

booktitle = "ICBBS 2022 - 2022 11th International Conference on Bioinformatics and Biomedical Science",

}

Zhang, Z, Ai, D, Geng, H & Yang, J 2022, CT-X-Ray Registration Via Spatial-Projective Dual Transformer Network Fused with Target Detection. in ICBBS 2022 - 2022 11th International Conference on Bioinformatics and Biomedical Science. ACM International Conference Proceeding Series, Association for Computing Machinery, pp. 93-98, 11th International Conference on Bioinformatics and Biomedical Science, ICBBS 2022, Nanning, China, 28/10/22. https://doi.org/10.1145/3571532.3571546

CT-X-Ray Registration Via Spatial-Projective Dual Transformer Network Fused with Target Detection. / Zhang, Zheng; Ai, Danni; Geng, Haixiao et al.
ICBBS 2022 - 2022 11th International Conference on Bioinformatics and Biomedical Science. Association for Computing Machinery, 2022. p. 93-98 (ACM International Conference Proceeding Series).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - CT-X-Ray Registration Via Spatial-Projective Dual Transformer Network Fused with Target Detection

AU - Zhang, Zheng

AU - Ai, Danni

AU - Geng, Haixiao

AU - Yang, Jian

PY - 2022/10/28

Y1 - 2022/10/28

N2 - Registration of CT-X-rays is crucial in high-precision orthopedic surgery. In this study, a deep learning network integrating convolution and transformer modules is proposed as a model for measuring image similarity for the registration of CT-X-rays. By training the network model to approximate the geodesic distance of Riemann space, the model has the property of convex function, to avoid falling into a local optimum. To further reduce the translation error of registration, this study introduces a spine detection network based on Yolov5, detects the spine of the target image and the image to be registered, obtains the spine position information and readjusts the translation component of the pose. The method used in this study has been tested, and the translation error and rotation error are lower than 3.05 mm and 1.96°, respectively.

AB - Registration of CT-X-rays is crucial in high-precision orthopedic surgery. In this study, a deep learning network integrating convolution and transformer modules is proposed as a model for measuring image similarity for the registration of CT-X-rays. By training the network model to approximate the geodesic distance of Riemann space, the model has the property of convex function, to avoid falling into a local optimum. To further reduce the translation error of registration, this study introduces a spine detection network based on Yolov5, detects the spine of the target image and the image to be registered, obtains the spine position information and readjusts the translation component of the pose. The method used in this study has been tested, and the translation error and rotation error are lower than 3.05 mm and 1.96°, respectively.

KW - CT-Xray registration

KW - Deep learning

KW - Spine detection

KW - Vision transformer

UR - http://www.scopus.com/inward/record.url?scp=85148444257&partnerID=8YFLogxK

U2 - 10.1145/3571532.3571546

DO - 10.1145/3571532.3571546

M3 - Conference contribution

AN - SCOPUS:85148444257

T3 - ACM International Conference Proceeding Series

SP - 93

EP - 98

BT - ICBBS 2022 - 2022 11th International Conference on Bioinformatics and Biomedical Science

PB - Association for Computing Machinery

T2 - 11th International Conference on Bioinformatics and Biomedical Science, ICBBS 2022

Y2 - 28 October 2022 through 30 October 2022

ER -

CT-X-Ray Registration Via Spatial-Projective Dual Transformer Network Fused with Target Detection

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this