Mutual information-based evaluation of image quality with its preliminary application to assessment of medical imaging systems

摘要

We describe an information-theoretic method for quanti-nfying overall image quality in terms of mutual information (MI). MI isnused to express the amount of information that an output imagencontains about an input object. The more the MI value provides, thenbetter the image quality is. Therefore, the overall quality of an imagencan be quantitatively evaluated by measuring MI. We demonstratednby way of image simulation that MI increases with increasing con-ntrast and decreases with the increase of noise and blur. We inves-ntigated the utility of this method by applying it to evaluate the per-nformance of four imaging plate detectors. We also comparednevaluation results in terms of MI against those in terms of the de-ntective quantum efficiency conventionally used for characterizing thenefficiency performance of imaging systems. Our results demonstratenthat the proposed method is simple to implement and has potentialnusefulness for evaluation of overall image quality.