ENRICHing NIH Imaging Datasets to Prepare them for Machine Learning

丰富 NIH 成像数据集，为机器学习做好准备

基本信息

批准号：
10842910
负责人：
Rima Arnaout
金额：
$ 35.09万
依托单位：
UNIVERSITY OF CALIFORNIA, SAN FRANCISCO
依托单位国家：
美国
项目类别：
财政年份：
2020
资助国家：
美国
起止时间：
2020-04-01 至 2025-03-31
项目状态：
未结题

来源：
https://reporter.nih.gov/project-details/10842910
关键词：
Achievement Algorithms Artificial Intelligence Benchmarking COVID-19 Cardiology Collaborations Collection Communities Data Data Science Data Set Data Storage and Retrieval Descriptor Detection Disease Ensure Environment Fetal Diseases Gender Goals Human Image Image Analysis Informatics Information Theory Intelligence Investments Label Learning Left Link Machine Learning Mathematics Measures Medical Imaging Modality Modeling Morphologic artifacts Paper Parents Pathology Patients Pattern Performance Physicians Process Property Proxy Publishing Race Research Research Methodology Research Personnel Roentgen Rays Running Scientist Techniques Testing Thoracic Radiography Time Training Ultrasonography United States National Institutes of Health Validation Work biomedical imaging clinical center computer program congenital heart disorder cost cost effective data harmonization data structure deep learning deep learning model disease classification diverse data frontier health care settings improved innovation insight large datasets multidisciplinary repository tool trustworthiness

项目摘要

PROJECT SUMMARY Objective: The goal of the parent proposal is to develop and optimize deep learning (DL) to improve detection of congenital heart disease (CHD) from fetal ultrasound imaging. This work includes evaluation of an imaging collection spanning two decades, tens of thousands of patients, and several clinical centers across a range of healthcare settings. Background: Through this work, we have found that performance of DL models is critically linked to the quality of the datasets used to train and test them. However, the AI/ML field lacks a complete understanding of how to measure “quality.” To date, image datasets are either described subjectively or measured crudely by size, i.e. the number of images they contain. However, “more is better” fails to account for the key importance of diversity in the quality of image datasets. In parent Aim 1, we sought to develop better metrics for dataset quality and content, founded in information theory and leveraging diversity. This work has already proven quite useful for our parent use case, but it is also extremely important for all imaging datasets in order to save on data storage/transfer costs, harmonize data intelligently, save on laborious image labeling, screen for artifacts both anticipated and un-anticipated, and ensure diversity at several levels. Preliminary Studies: Our multi-disciplinary team in imaging, DL, and information theory has successfully developed a framework to analyze image datasets, called ENRICH. ENRICH consists of two main steps. First, a similarity metric is calculated for all pairs of images in a given dataset, forming a matrix of pairwise-similarity values. Second, an instance-selection algorithm operates on the matrix to describe its diversity and/or curate the most informative images. ENRICH is customizable in that different choices for pairwise image similarity metric and for curation algorithm can be used for different tasks. An initial implementation of ENRICH aimed at reducing redundancy allowed us to get the same DL model performance in a CHD classification task from only a fraction of the original training data. It also identified data structure and imaging artifacts without a priori labeling, among other achievements (see Research Strategy). Goals of Supplement: The next logical step is to apply ENRICH to more biomedical datasets, both to further validate its utility and to provide quantitative descriptors of quality on datasets important for the research community. Aims: (1) We will run ENRICH on several NIH imaging datasets, including (2) validating labels and adding annotations to targeted subsets of these datasets. (3) We will document and publish these methods for the research community to use, including connecting with the original NIH repository for each dataset. Environment and Impact: This work proposed is supported in an outstanding environment at the crossroads of data science, imaging, and information theory and will provide valuable tools and insight into how best to measure image dataset content and quality in order to rigorously train and test DL for biomedical tasks.

项目摘要目的：父母建议的目标是开发和优化深度学习（DL）以改善检测胎儿超声成像的先天性心脏病（CHD）。这项工作包括对成像的评估跨越二十年的收集，成千上万的患者，以及多个临床中心医疗保健设置。背景：通过这项工作，我们发现DL模型的性能是与用于训练和测试的数据集的质量密切相关。但是，AI/ML场缺乏完全了解如何衡量“质量”。迄今为止，图像数据集的主观描述或用大小粗略测量，即它们包含的图像数量。但是，“更多更好”无法说明因为多样性在图像数据集质量中的关键重要性。在父母目标1中，我们试图发展在信息理论和利用多样性中建立的数据集质量和内容的更好的指标。这项工作已经证明对我们的父母用例很有用，但是对于所有成像也非常重要数据集为了节省数据存储/传输成本，智能地协调数据，保存在实验室图像上标记，屏幕构成预期和未孕期的雕像，并确保多样性在多个层面上。初步研究：我们的成像，DL和信息理论领域的多学科团队已成功开发了一个框架来分析图像数据集，称为富集。丰富由两个主要步骤组成。第一的，计算给定数据集中所有图像对的相似性度量，形成成对相似性的矩阵值。其次，实例选择算法在矩阵上运行，以描述其多样性和/或策展人最有用的图像。富集是可以自定义的，以对成对图像相似性的不同选择度量和用于策展算法可用于不同的任务。针对的最初实施降低冗余使我们仅从CHD分类任务中获得相同的DL模型性能原始培训数据的一小部分。它还在没有先验的情况下确定了数据结构和成像伪像标签，除其他成就（请参阅研究策略）。补充目标：下一个逻辑步骤是要应用于更多的生物医学数据集，以进一步验证其效用并提供定量数据集质量的描述对研究界很重要。目的：（1）我们将富裕几个NIH成像数据集，包括（2）验证标签并将注释添加到目标子集中这些数据集。（3）我们将记录并发布这些方法供研究界使用，包括与每个数据集的原始NIH存储库连接。环境与影响：提出的这项工作是在数据科学，成像和信息理论的十字路口的杰出环境中得到支持并将提供有价值的工具，并了解如何最好地测量图像数据集内容和质量严格训练和测试DL的生物医学任务。

项目成果

期刊论文数量（12）

专著数量（0）

科研奖励数量（0）

会议论文数量（0）

专利数量（0）

Visualizing omicron: COVID-19 deaths vs. cases over time.

DOI：
10.1371/journal.pone.0265233
发表时间：
2022
期刊：
PloS one
影响因子：
3.7
作者：
通讯作者：

The (Heart and) Soul of a Human Creation: Designing Echocardiography for the Big Data Age.

人类创造的（心和）灵魂：为大数据时代设计超声心动图。

DOI：
10.1016/j.echo.2023.04.016
发表时间：
2023
期刊：
Journal of the American Society of Echocardiography : official publication of the American Society of Echocardiography
影响因子：
0
作者：
Arnaout,Rima;Hahn,RebeccaT;Hung,JudyW;Jone,Pei-Ni;Lester,StevenJ;Little,StephenH;Mackensen,GBurkhard;Rigolin,Vera;Sachdev,Vandana;Saric,Muhamed;Sengupta,ParthoP;Strom,JordanB;Taub,CynthiaC;Thamman,Ritu;Abraham,Theodore
通讯作者：
Abraham,Theodore

Mitral Valve Atlas for Artificial Intelligence Predictions of MitraClip Intervention Outcomes.

DOI：
10.3389/fcvm.2021.759675
发表时间：
2021
期刊：
Frontiers in cardiovascular medicine
影响因子：
3.6
作者：
Dabiri Y;Yao J;Mahadevan VS;Gruber D;Arnaout R;Gentzsch W;Guccione JM;Kassab GS
通讯作者：
Kassab GS

An ensemble of neural networks provides expert-level prenatal detection of complex congenital heart disease.