Validation and Statistical Power Comparison of Methods for Analyzing Free-response Observer Performance Studies

Based on this work, the recommendation is to use JAFROC-1 for human observers (including human observers with CAD assist) and the NP method for evaluating CAD algorithms.