[Use of compressed images for medical checkups: comparison between radiologists' subjective quality assessment and diagnostic usefulness]

Abstract

With the increasing demand on picture archiving and communication systems (PACS) to handle larger numbers of images, much research has been conducted on the utilization of lossy compression for efficient communication and storage in PACS systems. The receiver operating characteristics (ROC) method is commonly used to determine the appropriate compression ratio of lossy images. However, conducting an ROC experiment is impractical at actual medical facilities because it is difficult to prepare and interpret images with different compression ratios for each modality, body part, and acquisition method. On the other hand, experienced radiologists have the ability to subjectively assess the level of image quality required for interpretatior and quantitative analysis. Therefore, by simply viewing images, they are able to determine the appropriate criteria for image quality and derive a practical compression ratio for each application. The present study focused on lossy compression as employed in teleradiology systems for medical checkups. Experiments comparing radiologists' subjective assessment of compressed images against the diagnostic results obtained using these images were conducted in order to investigate appropriate compression ratios and efficient methods for determining them.