Significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre-annotation.

This study investigates the feasibility of reducing manual image annotation costs in medical image database construction by utilizing a step by step approach where the Artificial Intelligence model (AI model) trained on a previous batch of data automatically pre-annotates the next batch of image dat...

Full description

Saved in:
Bibliographic Details
Main Authors: Fu Zheng, Liu XingMing, Xu JuYing, Tao MengYing, Yang BaoJian, Shan Yan, Ye KeWei, Lu ZhiKai, Huang Cheng, Qi KeLan, Chen XiHao, Du WenFei, He Ping, Wang RunYu, Ying Ying, Bu XiaoHui
Format: Article
Language:English
Published: Public Library of Science (PLoS) 2025-06-01
Series:PLOS Digital Health
Online Access:https://doi.org/10.1371/journal.pdig.0000738
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849427368843673600
author Fu Zheng
Liu XingMing
Xu JuYing
Tao MengYing
Yang BaoJian
Shan Yan
Ye KeWei
Lu ZhiKai
Huang Cheng
Qi KeLan
Chen XiHao
Du WenFei
He Ping
Wang RunYu
Ying Ying
Bu XiaoHui
author_facet Fu Zheng
Liu XingMing
Xu JuYing
Tao MengYing
Yang BaoJian
Shan Yan
Ye KeWei
Lu ZhiKai
Huang Cheng
Qi KeLan
Chen XiHao
Du WenFei
He Ping
Wang RunYu
Ying Ying
Bu XiaoHui
author_sort Fu Zheng
collection DOAJ
description This study investigates the feasibility of reducing manual image annotation costs in medical image database construction by utilizing a step by step approach where the Artificial Intelligence model (AI model) trained on a previous batch of data automatically pre-annotates the next batch of image data, taking ultrasound image of thyroid nodule annotation as an example. The study used YOLOv8 as the AI model. During the AI model training, in addition to conventional image augmentation techniques, augmentation methods specifically tailored for ultrasound images were employed to balance the quantity differences between thyroid nodule classes and enhance model training effectiveness. The study found that training the model with augmented data significantly outperformed training with raw images data. When the number of original images number was only 1,360, with 7 thyroid nodule classifications, pre-annotation using the AI model trained on augmented data could save at least 30% of the manual annotation workload for junior physicians. When the scale of original images number reached 6,800, the classification accuracy of the AI model trained on augmented data was very close with that of junior physicians, eliminating the need for manual preliminary annotation.
format Article
id doaj-art-88fa595e1a0648f5810c24f3d5a0cb39
institution Kabale University
issn 2767-3170
language English
publishDate 2025-06-01
publisher Public Library of Science (PLoS)
record_format Article
series PLOS Digital Health
spelling doaj-art-88fa595e1a0648f5810c24f3d5a0cb392025-08-20T03:29:03ZengPublic Library of Science (PLoS)PLOS Digital Health2767-31702025-06-0146e000073810.1371/journal.pdig.0000738Significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre-annotation.Fu ZhengLiu XingMingXu JuYingTao MengYingYang BaoJianShan YanYe KeWeiLu ZhiKaiHuang ChengQi KeLanChen XiHaoDu WenFeiHe PingWang RunYuYing YingBu XiaoHuiThis study investigates the feasibility of reducing manual image annotation costs in medical image database construction by utilizing a step by step approach where the Artificial Intelligence model (AI model) trained on a previous batch of data automatically pre-annotates the next batch of image data, taking ultrasound image of thyroid nodule annotation as an example. The study used YOLOv8 as the AI model. During the AI model training, in addition to conventional image augmentation techniques, augmentation methods specifically tailored for ultrasound images were employed to balance the quantity differences between thyroid nodule classes and enhance model training effectiveness. The study found that training the model with augmented data significantly outperformed training with raw images data. When the number of original images number was only 1,360, with 7 thyroid nodule classifications, pre-annotation using the AI model trained on augmented data could save at least 30% of the manual annotation workload for junior physicians. When the scale of original images number reached 6,800, the classification accuracy of the AI model trained on augmented data was very close with that of junior physicians, eliminating the need for manual preliminary annotation.https://doi.org/10.1371/journal.pdig.0000738
spellingShingle Fu Zheng
Liu XingMing
Xu JuYing
Tao MengYing
Yang BaoJian
Shan Yan
Ye KeWei
Lu ZhiKai
Huang Cheng
Qi KeLan
Chen XiHao
Du WenFei
He Ping
Wang RunYu
Ying Ying
Bu XiaoHui
Significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre-annotation.
PLOS Digital Health
title Significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre-annotation.
title_full Significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre-annotation.
title_fullStr Significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre-annotation.
title_full_unstemmed Significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre-annotation.
title_short Significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre-annotation.
title_sort significant reduction in manual annotation costs in ultrasound medical image database construction through step by step artificial intelligence pre annotation
url https://doi.org/10.1371/journal.pdig.0000738
work_keys_str_mv AT fuzheng significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT liuxingming significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT xujuying significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT taomengying significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT yangbaojian significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT shanyan significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT yekewei significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT luzhikai significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT huangcheng significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT qikelan significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT chenxihao significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT duwenfei significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT heping significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT wangrunyu significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT yingying significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation
AT buxiaohui significantreductioninmanualannotationcostsinultrasoundmedicalimagedatabaseconstructionthroughstepbystepartificialintelligencepreannotation