The providers' Graticule Life Image Machine Parsed Set (GLIMPS) consists of patient imaging data deidentified using a machine-learning algorithm and a summary of features generated from the data using natural language processing. The data repository includes 92 million deidentified diagnostic images from 140,000 patients throughout the U.S.
GLIMPS provides users with additional layers of clinical context to deidentified data, such as key fields from DICOM headers, and allows users to extract measurable values such as ejection fraction from echocardiograms. The dataset is also compatible with the Observational Health Data Sciences and Informatics (OHDSI) common data model used by many life sciences companies and regulators.
GLIMPS can be used for several medical applications, including the following:
- Validating AI models on organized clinical data
- Providing data to support the development of new techniques
- Allowing AI researchers to distinguish control data from trial data
- Identifying novel imaging biomarkers for rare or new disease subtypes
- Identifying disease markers before diagnosis and examining disease progression during treatment
Copyright © 2019 AuntMinnie.com