One of the key limitations of existing image processing algorithms for computer-aided detection (CADe) is that they are often designed and evaluated in an ad hoc manner.

This paper characterizes some of the issues and shortcomings in existing performance evaluation paradigms for image processing algorithms in breast cancer screening, particularly in the context of computer aided detection. We present the framework for establishing a performance evaluation process using standardized criteria. We conclude with some specific recommendations to improve the infrastructure for evaluation the performance of image processing algorithms.

Key words: Mammogram, Algorithm, Design process, Performance evaluation, Image databases.