Deep feature descriptor based hierarchical dense matching for X-ray angiographic images

Jingfan Fan, Jian Yang, Yachen Wang, Siyuan Yang, Danni Ai, Yong Huang, Hong Song, Yongtian Wang, Dinggang Shen

Research output: Contribution to journalArticle

Abstract

Backgroud and Objective: X-ray angiography, a powerful technique for blood vessel visualization, is widely used for interventional diagnosis of coronary artery disease because of its fast imaging speed and perspective inspection ability. Matching feature points in angiographic images is a considerably challenging task due to repetitive weak-textured regions. Methods: In this paper, we propose an angiographic image matching method based on the hierarchical dense matching framework, where a novel deep feature descriptor is designed to compute multilevel correlation maps. In particular, the deep feature descriptor is computed by a deep learning model specifically designed and trained for angiographic images, thereby making the correlation maps more distinctive for corresponding feature points in different angiographic images. Moreover, point correspondences are further hierarchically extracted from multilevel correlation maps with the highest similarity response(s), which is relatively robust and accurate. To overcome the problem regarding the lack of training samples, the convolutional neural network (designed for deep feature descriptor) is initially trained on samples from natural images and then fine-tuned on manually annotated angiographic images. Finally, a dense matching completion method, based on the distance between deep feature descriptors, is proposed to generate dense matches between images. Results: The proposed method has been evaluated on the number and accuracy of extracted matches and the performance of subtraction images. Experiments on a variety of angiographic images show promising matching accuracy, compared with state-of-the-art methods. Conclusions: The proposed angiographic image matching method is shown to be accurate and effective for feature matching in angiographic images, and further achieves good performance in image subtraction.

Original languageEnglish
Pages (from-to)233-242
Number of pages10
JournalComputer Methods and Programs in Biomedicine
Volume175
DOIs
Publication statusPublished - 2019 Jul 1

Fingerprint

Image matching
X-Rays
X rays
Angiography
Blood vessels
Visualization
Inspection
Neural networks
Imaging techniques
Aptitude
Blood Vessels
Coronary Artery Disease
Experiments
Learning
Deep learning

Keywords

  • Convolutional neural network
  • Coronary artery
  • Hierarchical dense matching

ASJC Scopus subject areas

  • Software
  • Computer Science Applications
  • Health Informatics

Cite this

Deep feature descriptor based hierarchical dense matching for X-ray angiographic images. / Fan, Jingfan; Yang, Jian; Wang, Yachen; Yang, Siyuan; Ai, Danni; Huang, Yong; Song, Hong; Wang, Yongtian; Shen, Dinggang.

In: Computer Methods and Programs in Biomedicine, Vol. 175, 01.07.2019, p. 233-242.

Research output: Contribution to journalArticle

Fan, Jingfan ; Yang, Jian ; Wang, Yachen ; Yang, Siyuan ; Ai, Danni ; Huang, Yong ; Song, Hong ; Wang, Yongtian ; Shen, Dinggang. / Deep feature descriptor based hierarchical dense matching for X-ray angiographic images. In: Computer Methods and Programs in Biomedicine. 2019 ; Vol. 175. pp. 233-242.
@article{62f2fb0683f2428ab782aef221c78f3b,
title = "Deep feature descriptor based hierarchical dense matching for X-ray angiographic images",
abstract = "Backgroud and Objective: X-ray angiography, a powerful technique for blood vessel visualization, is widely used for interventional diagnosis of coronary artery disease because of its fast imaging speed and perspective inspection ability. Matching feature points in angiographic images is a considerably challenging task due to repetitive weak-textured regions. Methods: In this paper, we propose an angiographic image matching method based on the hierarchical dense matching framework, where a novel deep feature descriptor is designed to compute multilevel correlation maps. In particular, the deep feature descriptor is computed by a deep learning model specifically designed and trained for angiographic images, thereby making the correlation maps more distinctive for corresponding feature points in different angiographic images. Moreover, point correspondences are further hierarchically extracted from multilevel correlation maps with the highest similarity response(s), which is relatively robust and accurate. To overcome the problem regarding the lack of training samples, the convolutional neural network (designed for deep feature descriptor) is initially trained on samples from natural images and then fine-tuned on manually annotated angiographic images. Finally, a dense matching completion method, based on the distance between deep feature descriptors, is proposed to generate dense matches between images. Results: The proposed method has been evaluated on the number and accuracy of extracted matches and the performance of subtraction images. Experiments on a variety of angiographic images show promising matching accuracy, compared with state-of-the-art methods. Conclusions: The proposed angiographic image matching method is shown to be accurate and effective for feature matching in angiographic images, and further achieves good performance in image subtraction.",
keywords = "Convolutional neural network, Coronary artery, Hierarchical dense matching",
author = "Jingfan Fan and Jian Yang and Yachen Wang and Siyuan Yang and Danni Ai and Yong Huang and Hong Song and Yongtian Wang and Dinggang Shen",
year = "2019",
month = "7",
day = "1",
doi = "10.1016/j.cmpb.2019.04.006",
language = "English",
volume = "175",
pages = "233--242",
journal = "Computer Methods and Programs in Biomedicine",
issn = "0169-2607",
publisher = "Elsevier Ireland Ltd",

}

TY - JOUR

T1 - Deep feature descriptor based hierarchical dense matching for X-ray angiographic images

AU - Fan, Jingfan

AU - Yang, Jian

AU - Wang, Yachen

AU - Yang, Siyuan

AU - Ai, Danni

AU - Huang, Yong

AU - Song, Hong

AU - Wang, Yongtian

AU - Shen, Dinggang

PY - 2019/7/1

Y1 - 2019/7/1

N2 - Backgroud and Objective: X-ray angiography, a powerful technique for blood vessel visualization, is widely used for interventional diagnosis of coronary artery disease because of its fast imaging speed and perspective inspection ability. Matching feature points in angiographic images is a considerably challenging task due to repetitive weak-textured regions. Methods: In this paper, we propose an angiographic image matching method based on the hierarchical dense matching framework, where a novel deep feature descriptor is designed to compute multilevel correlation maps. In particular, the deep feature descriptor is computed by a deep learning model specifically designed and trained for angiographic images, thereby making the correlation maps more distinctive for corresponding feature points in different angiographic images. Moreover, point correspondences are further hierarchically extracted from multilevel correlation maps with the highest similarity response(s), which is relatively robust and accurate. To overcome the problem regarding the lack of training samples, the convolutional neural network (designed for deep feature descriptor) is initially trained on samples from natural images and then fine-tuned on manually annotated angiographic images. Finally, a dense matching completion method, based on the distance between deep feature descriptors, is proposed to generate dense matches between images. Results: The proposed method has been evaluated on the number and accuracy of extracted matches and the performance of subtraction images. Experiments on a variety of angiographic images show promising matching accuracy, compared with state-of-the-art methods. Conclusions: The proposed angiographic image matching method is shown to be accurate and effective for feature matching in angiographic images, and further achieves good performance in image subtraction.

AB - Backgroud and Objective: X-ray angiography, a powerful technique for blood vessel visualization, is widely used for interventional diagnosis of coronary artery disease because of its fast imaging speed and perspective inspection ability. Matching feature points in angiographic images is a considerably challenging task due to repetitive weak-textured regions. Methods: In this paper, we propose an angiographic image matching method based on the hierarchical dense matching framework, where a novel deep feature descriptor is designed to compute multilevel correlation maps. In particular, the deep feature descriptor is computed by a deep learning model specifically designed and trained for angiographic images, thereby making the correlation maps more distinctive for corresponding feature points in different angiographic images. Moreover, point correspondences are further hierarchically extracted from multilevel correlation maps with the highest similarity response(s), which is relatively robust and accurate. To overcome the problem regarding the lack of training samples, the convolutional neural network (designed for deep feature descriptor) is initially trained on samples from natural images and then fine-tuned on manually annotated angiographic images. Finally, a dense matching completion method, based on the distance between deep feature descriptors, is proposed to generate dense matches between images. Results: The proposed method has been evaluated on the number and accuracy of extracted matches and the performance of subtraction images. Experiments on a variety of angiographic images show promising matching accuracy, compared with state-of-the-art methods. Conclusions: The proposed angiographic image matching method is shown to be accurate and effective for feature matching in angiographic images, and further achieves good performance in image subtraction.

KW - Convolutional neural network

KW - Coronary artery

KW - Hierarchical dense matching

UR - http://www.scopus.com/inward/record.url?scp=85064888961&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064888961&partnerID=8YFLogxK

U2 - 10.1016/j.cmpb.2019.04.006

DO - 10.1016/j.cmpb.2019.04.006

M3 - Article

VL - 175

SP - 233

EP - 242

JO - Computer Methods and Programs in Biomedicine

JF - Computer Methods and Programs in Biomedicine

SN - 0169-2607

ER -