Berkeley AI Research proposes information-driven design for imaging systems
Berkeley AI Research introduced a framework that estimates mutual information between objects and measurements to evaluate and optimize imaging systems. The method computes H(Y|X) directly from known noise physics and le...