Image feature index: A novel metric for quantifying chest radiographic image quality

Author:

Liang Zejun1,Tang Jing1,Xu Peng2,Zeng Wen1,Zhang Jiye2,Zhang Yu1,Zeng Lingming1,Wang Hairong1,Xia Chunchao1,Li Zhenlin1

Affiliation:

1. Department of Radiology, West China Hospital Sichuan University Chengdu China

2. Radiology Department Shenzhen Mindray Bio‐Medical Electronics Co., Ltd Shenzhen China

Abstract

AbstractBackgroundRadiographic X‐ray imaging is a common clinical examination. Current objective methods for quantifying image quality for radiographs struggle to capture the combined impact of factors throughout the imaging chain on the perceived image quality. Therefore, there is a need to further develop metrics that correlate with image quality as perceived by the observer.ObjectivesWe proposed the image feature index (IFI) to comprehensively quantify radiographic X‐ray image quality. We also aimed to study the correlation between IFI and observer‐perceived image quality for chest radiographs.Materials and methodsThe IFI algorithm was developed, which measured the amount of information, textural features, and noise in the image. A total of 70 chest phantom radiographs were generated under 60–120 kV and 0.2–80 mAs. A vendor‐proprietary exposure index (EI) and dose area product (DAP) were extracted from the DICOM header in addition to calculating IFI for each image to investigate the relationships between IFI, EI and DAP. The quality of the images was rated by three observers, and the correlation between IFI and subjective score of image quality was tested. Next, a retrospective study using a random sample of 50 clinical chest radiographs was performed, and the correlation between IFI and subjective score was tested. The correlation was determined by the Spearman test.ResultsThe curves of IFI versus DAP and IFI versus EI both demonstrated a similar three‐stage form where IFI is above zero: in the first stage, IFI increases rapidly with increased DAP or EI, whereas in the second stage, the slope of the curves decreased towards an asymptote, that is, minimal gain in IFI with increased DAP or EI—until they hit the inflection point and then descended sharply in the third stage. For both phantom and clinical chest images, IFI demonstrated good correlation with subjective score (r = 0.9084 for phantom images, r = 0.8153 for clinical images).ConclusionsIFI is a feasible and efficient descriptor for image quality for chest radiographs. Future studies with larger sample sizes and sample types are needed to confirm the feasibility of IFI for other exam types and anatomical views, thus fulfilling and extending the potential applications of IFI in quality control and radiation dose reduction.

Funder

West China Hospital, Sichuan University

Publisher

Wiley

Subject

General Medicine

全球学者库

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"全球学者库"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前全球学者库共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2023 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3