Image Quality and Task Complexity for Machine Learning

R-DEX Systems, Inc. and its strategic partner will combine their joint expertise in information theory, machine learning, and image analysis to develop an information-theoretic framework linking task complexity to image quality and model capacity for image discrimination tasks. The team will perform a comprehensive literature review of state-of-the-art techniques for task complexity and image quality analysis, develop a prototype model based on information theory, and apply it to a variety of situations with electro-optical imagery. The team will then extend the framework to unlabeled conditions to assess image quality and data complexity and will design a unified approach that supports automation within machine learning workflows. Finally, the team will develop a test plan for transitioning the Phase I prototype to a Phase II development effort.