Skip to main content
. 2023 Jan 31;2(1):e0000189. doi: 10.1371/journal.pdig.0000189

Table 2. Reported (expected) strengths and promises of image-based machine learning.

Domain Subdomain(s) Descriptions
Analytic power accuracy ability to classify images as well as experts, ensuring diagnostic and prognostic accuracy
superiority ability to differentiate images better than an expert, identifying patterns not always visible to the human eye
objectivity lack of human subjective biases and errors, reducing variability and improving comparability and reproducibility
big data use ability to handle and analyse large amounts of data and tackle the challenges of big data
Efficiency time efficiency ability to speed up analysis and clinical translation through automation of otherwise time-consuming manual tasks
cost efficiency ability to lower direct and indirect costs through time and diagnostic efficiency, automation, and enhanced workflows
Clinical impact workflow improvements ability to optimize clinical workflows through integrating, automating, streamlining, and structuring processes
decision making supporting faster, cheaper, more accurate and higher-level decision making and clinical interpretations
personalization facilitate personalized care, though higher analytic power, efficiency and improved clinical workflows
Equity reach, access, and affordability promise of increased geographic reach of and better access to affordable image-based healthcare