An efficient annotated data generation method for echocardiographic image segmentation
Copyright © 2022 Elsevier Ltd. All rights reserved..
BACKGROUND: In recent years, deep learning techniques have demonstrated promising performances in echocardiography (echo) data segmentation, which constitutes a critical step in the diagnosis and prognosis of cardiovascular diseases (CVDs). However, their successful implementation requires large number and high-quality annotated samples, whose acquisition is arduous and expertise-demanding. To this end, this study aims at circumventing the tedious, time-consuming and expertise-demanding data annotation involved in deep learning-based echo data segmentation.
METHODS: We propose a two-phase framework for fast generation of annotated echo data needed for implementing intelligent cardiac structure segmentation systems. First, multi-size and multi-orientation cardiac structures are simulated leveraging polynomial fitting method. Second, the obtained cardiac structures are embedded onto curated endoscopic ultrasound images using Fourier Transform algorithm, resulting in pairs of annotated samples. The practical significance of the proposed framework is validated through using the generated realistic annotated images as auxiliary dataset to pretrain deep learning models for automatic segmentation of left ventricle and left ventricle wall in real echo data, respectively.
RESULTS: Extensive experimental analyses indicate that compared with training from scratch, fine-tuning after pretraining with the generated dataset always results in significant performance improvement whereby the improvement margins in terms of Dice and IoU can reach 12.9% and 7.74%, respectively.
CONCLUSION: The proposed framework has great potential to overcome the shortage of labeled data hampering the deployment of deep learning approaches in echo data analysis.
| Medienart: |
E-Artikel |
|---|
| Erscheinungsjahr: |
2022 |
|---|---|
| Erschienen: |
2022 |
| Enthalten in: |
Zur Gesamtaufnahme - volume:149 |
|---|---|
| Enthalten in: |
Computers in biology and medicine - 149(2022) vom: 20. Okt., Seite 106090 |
| Sprache: |
Englisch |
|---|
| Beteiligte Personen: |
Monkam, Patrice [VerfasserIn] |
|---|
| Links: |
|---|
| Themen: |
Annotated data generation |
|---|
| Anmerkungen: |
Date Completed 23.09.2022 Date Revised 30.09.2022 published: Print-Electronic Citation Status MEDLINE |
|---|
| doi: |
10.1016/j.compbiomed.2022.106090 |
|---|
| funding: |
|
|---|---|
| Förderinstitution / Projekttitel: |
|
| PPN (Katalog-ID): |
NLM346320321 |
|---|
| LEADER | 01000naa a22002652 4500 | ||
|---|---|---|---|
| 001 | NLM346320321 | ||
| 003 | DE-627 | ||
| 005 | 20231226031144.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 231226s2022 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1016/j.compbiomed.2022.106090 |2 doi | |
| 028 | 5 | 2 | |a pubmed24n1154.xml |
| 035 | |a (DE-627)NLM346320321 | ||
| 035 | |a (NLM)36115304 | ||
| 035 | |a (PII)S0010-4825(22)00798-3 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng | ||
| 100 | 1 | |a Monkam, Patrice |e verfasserin |4 aut | |
| 245 | 1 | 3 | |a An efficient annotated data generation method for echocardiographic image segmentation |
| 264 | 1 | |c 2022 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a ƒaComputermedien |b c |2 rdamedia | ||
| 338 | |a ƒa Online-Ressource |b cr |2 rdacarrier | ||
| 500 | |a Date Completed 23.09.2022 | ||
| 500 | |a Date Revised 30.09.2022 | ||
| 500 | |a published: Print-Electronic | ||
| 500 | |a Citation Status MEDLINE | ||
| 520 | |a Copyright © 2022 Elsevier Ltd. All rights reserved. | ||
| 520 | |a BACKGROUND: In recent years, deep learning techniques have demonstrated promising performances in echocardiography (echo) data segmentation, which constitutes a critical step in the diagnosis and prognosis of cardiovascular diseases (CVDs). However, their successful implementation requires large number and high-quality annotated samples, whose acquisition is arduous and expertise-demanding. To this end, this study aims at circumventing the tedious, time-consuming and expertise-demanding data annotation involved in deep learning-based echo data segmentation | ||
| 520 | |a METHODS: We propose a two-phase framework for fast generation of annotated echo data needed for implementing intelligent cardiac structure segmentation systems. First, multi-size and multi-orientation cardiac structures are simulated leveraging polynomial fitting method. Second, the obtained cardiac structures are embedded onto curated endoscopic ultrasound images using Fourier Transform algorithm, resulting in pairs of annotated samples. The practical significance of the proposed framework is validated through using the generated realistic annotated images as auxiliary dataset to pretrain deep learning models for automatic segmentation of left ventricle and left ventricle wall in real echo data, respectively | ||
| 520 | |a RESULTS: Extensive experimental analyses indicate that compared with training from scratch, fine-tuning after pretraining with the generated dataset always results in significant performance improvement whereby the improvement margins in terms of Dice and IoU can reach 12.9% and 7.74%, respectively | ||
| 520 | |a CONCLUSION: The proposed framework has great potential to overcome the shortage of labeled data hampering the deployment of deep learning approaches in echo data analysis | ||
| 650 | 4 | |a Journal Article | |
| 650 | 4 | |a Research Support, Non-U.S. Gov't | |
| 650 | 4 | |a Annotated data generation | |
| 650 | 4 | |a Cardiac structure segmentation | |
| 650 | 4 | |a Echocardiography | |
| 650 | 4 | |a Performance improvement | |
| 650 | 4 | |a Transfer learning | |
| 700 | 1 | |a Jin, Songbai |e verfasserin |4 aut | |
| 700 | 1 | |a Lu, Wenkai |e verfasserin |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t Computers in biology and medicine |d 1970 |g 149(2022) vom: 20. Okt., Seite 106090 |w (DE-627)NLM000382272 |x 1879-0534 |7 nnns |
| 773 | 1 | 8 | |g volume:149 |g year:2022 |g day:20 |g month:10 |g pages:106090 |
| 856 | 4 | 0 | |u http://dx.doi.org/10.1016/j.compbiomed.2022.106090 |3 Volltext |
| 912 | |a GBV_USEFLAG_A | ||
| 912 | |a GBV_NLM | ||
| 951 | |a AR | ||
| 952 | |d 149 |j 2022 |b 20 |c 10 |h 106090 | ||