As life science organizations implement more artificial intelligence and machine learning in their R&D processes, they rely on the services of expert readers to label and annotate their medical imaging assets for downstream analysis. But administering reader workflows - assigning images to readers, adjudicating the resulting data and keeping this work efficient and compliant - is no simple feat, particularly at the scale necessary for machine learning. To date, research organizations have assembled various approaches for tackling this work. Many common tools used in reader studies are built for only a few rigid purposes, requiring researchers to invest significant upfront time creating data-capture systems external to the viewer. Once the data-capture system is built, readers must be trained to ensure accurate and adequate readings. Maintaining compliance is also a vital concern when research teams are collaborating on medical imaging. To perform these studies at scale, researchers should be able to design their own custom workflows without coding knowledge and with the confidence they are staying in compliance. Readers would be guided through their labeling workflow inside the viewer in a way that is intuitive and instructional while also providing beneficial data quality validation steps. Once reader data is collected, research teams then need additional tools to assist in adjudication, data standardization and subsequent research workflows.

Read on xtalks.com

Continue reading at xtalks.com

Medical Imaging Workflow

Facebook Share

Twitter Share

Linkedin Share

Email Share