LRTHR-Net: A Low-Resolution-to-High-Resolution Framework to Iteratively Refine the Segmentation of Thyroid Nodule in Ultrasound Images

Huai Chen; Shaoli Song; Xiuying Wang; Renzhen Wang; Deyu Meng; Lisheng Wang

doi:10.1007/978-3-030-71827-5_15

LRTHR-Net: A Low-Resolution-to-High-Resolution Framework to Iteratively Refine the Segmentation of Thyroid Nodule in Ultrasound Images

Huai Chen
, Shaoli Song
, Xiuying Wang
, Renzhen Wang
, Deyu Meng
, Lisheng Wang

School of Mathematics and Statistics

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

5 Scopus citations

Abstract

The thyroid nodule is quickly increasing worldwide and the thyroid ultrasound is the key tool for the diagnosis of it. For the subtle difference between malignant and benign nodules, segmenting lesions is the crucial preliminary step for diagnosis. In this paper, we propose a low-resolution-to-high-resolution segmentation framework for TN-SCUI2020 challenge to alleviate the workload of clinicians and improve the efficiency of diagnosis. Specifically speaking, in order to integrate multi-scale information, several low-resolution segmenting results are obtained firstly and combined with a high-resolution image to refine them and obtain high-resolution results. Secondly, iterative-transfer is proposed to effectively initialize network based on previous trained one on small-scale images. Finally, ensemble refinement is introduced to utilize multiple models to refine the segmentation again. Experimental results showed the effectiveness of the proposed framework. And we won the 2nd place in the segmentation task of TN-SCUI2020.

Original language	English
Title of host publication	Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data - MICCAI 2020 Challenges, ABCs 2020, L2R 2020, TN-SCUI 2020, Held in Conjunction with MICCAI 2020, Proceedings
Editors	Nadya Shusharina, Mattias P. Heinrich, Ruobing Huang
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	116-121
Number of pages	6
ISBN (Print)	9783030718268
DOIs	https://doi.org/10.1007/978-3-030-71827-5_15
State	Published - 2021
Event	Anatomical Brain Barriers to Cancer Spread: Segmentation from CT and MR Images Challenge, ABCs 2020, Learn2Reg Challenge, L2R 2020 and Thyroid Nodule Segmentation and Classification in Ultrasound Images Challenge, TN-SCUI 2020 held in conjunction with 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020 - Lima, Peru Duration: 4 Oct 2020 → 8 Oct 2020

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	12587 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	Anatomical Brain Barriers to Cancer Spread: Segmentation from CT and MR Images Challenge, ABCs 2020, Learn2Reg Challenge, L2R 2020 and Thyroid Nodule Segmentation and Classification in Ultrasound Images Challenge, TN-SCUI 2020 held in conjunction with 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020
Country/Territory	Peru
City	Lima
Period	4/10/20 → 8/10/20

Keywords

Iterative refinement
Low-resolution-to-high-resolution
TN-SCUI2020
Thyroid nodule segmentation
Thyroid ultrasound

Access to Document

10.1007/978-3-030-71827-5_15

Cite this

Chen, H., Song, S., Wang, X., Wang, R., Meng, D., & Wang, L. (2021). LRTHR-Net: A Low-Resolution-to-High-Resolution Framework to Iteratively Refine the Segmentation of Thyroid Nodule in Ultrasound Images. In N. Shusharina, M. P. Heinrich, & R. Huang (Eds.), Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data - MICCAI 2020 Challenges, ABCs 2020, L2R 2020, TN-SCUI 2020, Held in Conjunction with MICCAI 2020, Proceedings (pp. 116-121). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12587 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-71827-5_15

Chen, Huai ; Song, Shaoli ; Wang, Xiuying et al. / LRTHR-Net : A Low-Resolution-to-High-Resolution Framework to Iteratively Refine the Segmentation of Thyroid Nodule in Ultrasound Images. Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data - MICCAI 2020 Challenges, ABCs 2020, L2R 2020, TN-SCUI 2020, Held in Conjunction with MICCAI 2020, Proceedings. editor / Nadya Shusharina ; Mattias P. Heinrich ; Ruobing Huang. Springer Science and Business Media Deutschland GmbH, 2021. pp. 116-121 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{cbfa8a03ac2548c18553c4ddd61d56ff,

title = "LRTHR-Net: A Low-Resolution-to-High-Resolution Framework to Iteratively Refine the Segmentation of Thyroid Nodule in Ultrasound Images",

abstract = "The thyroid nodule is quickly increasing worldwide and the thyroid ultrasound is the key tool for the diagnosis of it. For the subtle difference between malignant and benign nodules, segmenting lesions is the crucial preliminary step for diagnosis. In this paper, we propose a low-resolution-to-high-resolution segmentation framework for TN-SCUI2020 challenge to alleviate the workload of clinicians and improve the efficiency of diagnosis. Specifically speaking, in order to integrate multi-scale information, several low-resolution segmenting results are obtained firstly and combined with a high-resolution image to refine them and obtain high-resolution results. Secondly, iterative-transfer is proposed to effectively initialize network based on previous trained one on small-scale images. Finally, ensemble refinement is introduced to utilize multiple models to refine the segmentation again. Experimental results showed the effectiveness of the proposed framework. And we won the 2nd place in the segmentation task of TN-SCUI2020.",

keywords = "Iterative refinement, Low-resolution-to-high-resolution, TN-SCUI2020, Thyroid nodule segmentation, Thyroid ultrasound",

author = "Huai Chen and Shaoli Song and Xiuying Wang and Renzhen Wang and Deyu Meng and Lisheng Wang",

note = "Publisher Copyright: {\textcopyright} 2021, Springer Nature Switzerland AG.; Anatomical Brain Barriers to Cancer Spread: Segmentation from CT and MR Images Challenge, ABCs 2020, Learn2Reg Challenge, L2R 2020 and Thyroid Nodule Segmentation and Classification in Ultrasound Images Challenge, TN-SCUI 2020 held in conjunction with 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020 ; Conference date: 04-10-2020 Through 08-10-2020",

year = "2021",

doi = "10.1007/978-3-030-71827-5\_15",

language = "英语",

isbn = "9783030718268",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "116--121",

editor = "Nadya Shusharina and Heinrich, \{Mattias P.\} and Ruobing Huang",

booktitle = "Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data - MICCAI 2020 Challenges, ABCs 2020, L2R 2020, TN-SCUI 2020, Held in Conjunction with MICCAI 2020, Proceedings",

}

Chen, H, Song, S, Wang, X, Wang, R, Meng, D & Wang, L 2021, LRTHR-Net: A Low-Resolution-to-High-Resolution Framework to Iteratively Refine the Segmentation of Thyroid Nodule in Ultrasound Images. in N Shusharina, MP Heinrich & R Huang (eds), Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data - MICCAI 2020 Challenges, ABCs 2020, L2R 2020, TN-SCUI 2020, Held in Conjunction with MICCAI 2020, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 12587 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 116-121, Anatomical Brain Barriers to Cancer Spread: Segmentation from CT and MR Images Challenge, ABCs 2020, Learn2Reg Challenge, L2R 2020 and Thyroid Nodule Segmentation and Classification in Ultrasound Images Challenge, TN-SCUI 2020 held in conjunction with 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020, Lima, Peru, 4/10/20. https://doi.org/10.1007/978-3-030-71827-5_15

LRTHR-Net: A Low-Resolution-to-High-Resolution Framework to Iteratively Refine the Segmentation of Thyroid Nodule in Ultrasound Images. / Chen, Huai; Song, Shaoli; Wang, Xiuying et al.
Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data - MICCAI 2020 Challenges, ABCs 2020, L2R 2020, TN-SCUI 2020, Held in Conjunction with MICCAI 2020, Proceedings. ed. / Nadya Shusharina; Mattias P. Heinrich; Ruobing Huang. Springer Science and Business Media Deutschland GmbH, 2021. p. 116-121 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 12587 LNCS).

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

TY - GEN

T1 - LRTHR-Net

T2 - Anatomical Brain Barriers to Cancer Spread: Segmentation from CT and MR Images Challenge, ABCs 2020, Learn2Reg Challenge, L2R 2020 and Thyroid Nodule Segmentation and Classification in Ultrasound Images Challenge, TN-SCUI 2020 held in conjunction with 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020

AU - Chen, Huai

AU - Song, Shaoli

AU - Wang, Xiuying

AU - Wang, Renzhen

AU - Meng, Deyu

AU - Wang, Lisheng

PY - 2021

Y1 - 2021

N2 - The thyroid nodule is quickly increasing worldwide and the thyroid ultrasound is the key tool for the diagnosis of it. For the subtle difference between malignant and benign nodules, segmenting lesions is the crucial preliminary step for diagnosis. In this paper, we propose a low-resolution-to-high-resolution segmentation framework for TN-SCUI2020 challenge to alleviate the workload of clinicians and improve the efficiency of diagnosis. Specifically speaking, in order to integrate multi-scale information, several low-resolution segmenting results are obtained firstly and combined with a high-resolution image to refine them and obtain high-resolution results. Secondly, iterative-transfer is proposed to effectively initialize network based on previous trained one on small-scale images. Finally, ensemble refinement is introduced to utilize multiple models to refine the segmentation again. Experimental results showed the effectiveness of the proposed framework. And we won the 2nd place in the segmentation task of TN-SCUI2020.

AB - The thyroid nodule is quickly increasing worldwide and the thyroid ultrasound is the key tool for the diagnosis of it. For the subtle difference between malignant and benign nodules, segmenting lesions is the crucial preliminary step for diagnosis. In this paper, we propose a low-resolution-to-high-resolution segmentation framework for TN-SCUI2020 challenge to alleviate the workload of clinicians and improve the efficiency of diagnosis. Specifically speaking, in order to integrate multi-scale information, several low-resolution segmenting results are obtained firstly and combined with a high-resolution image to refine them and obtain high-resolution results. Secondly, iterative-transfer is proposed to effectively initialize network based on previous trained one on small-scale images. Finally, ensemble refinement is introduced to utilize multiple models to refine the segmentation again. Experimental results showed the effectiveness of the proposed framework. And we won the 2nd place in the segmentation task of TN-SCUI2020.

KW - Iterative refinement

KW - Low-resolution-to-high-resolution

KW - TN-SCUI2020

KW - Thyroid nodule segmentation

KW - Thyroid ultrasound

UR - https://www.scopus.com/pages/publications/85104490744

U2 - 10.1007/978-3-030-71827-5_15

DO - 10.1007/978-3-030-71827-5_15

M3 - 会议稿件

AN - SCOPUS:85104490744

SN - 9783030718268

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 116

EP - 121

BT - Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data - MICCAI 2020 Challenges, ABCs 2020, L2R 2020, TN-SCUI 2020, Held in Conjunction with MICCAI 2020, Proceedings

A2 - Shusharina, Nadya

A2 - Heinrich, Mattias P.

A2 - Huang, Ruobing

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 4 October 2020 through 8 October 2020

ER -

Chen H, Song S, Wang X, Wang R, Meng D, Wang L. LRTHR-Net: A Low-Resolution-to-High-Resolution Framework to Iteratively Refine the Segmentation of Thyroid Nodule in Ultrasound Images. In Shusharina N, Heinrich MP, Huang R, editors, Segmentation, Classification, and Registration of Multi-modality Medical Imaging Data - MICCAI 2020 Challenges, ABCs 2020, L2R 2020, TN-SCUI 2020, Held in Conjunction with MICCAI 2020, Proceedings. Springer Science and Business Media Deutschland GmbH. 2021. p. 116-121. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-71827-5_15