Study on Optimization of Archiving Condition Under Jpeg2000 and Its Quality Evaluation for Medical Ultrasonic Echo Image

Abstract

Medical ultrasonic echo image is indispensable diagnostic technique and a large number of the investigation is carried out every day. Therefore, the image compression technique is very important. Author has proposed a technique using Genetic Algorithm (GA) to optimize quantization condition in JPEG2000. The result has shown image quality evaluation by PSNR (the objective assessment) and one by ultrasonographer (the subjective assessment by PSNR) are sometimes not equivalent. In this research, some objective assessment methods and the subjective assessment are compared for 4 images whose diseases part and scanning method are different (image group A). The result shows the objective assessment by a contrast is more consistent with the subjective assessment than the objective assessment by PSNR. Next, proposed GA is modified by using the contrast factor and image group A to optimize an archiving condition. The compressed images by the modified GA are evaluated by the objective assessment by the contrast and the subjective assessment. In this experiment, other 4 images, which are not used to decide the archiving condition, are also used (image group B). As the result, the archiving condition is effective for not only image group A but also image group B. That means the results of the objective assessment by the contrast and the subjective assessment are equivalent whenever using the archiving condition optimized by the modified GA using some typical ultrasonic images. The result also suggests the objective assessment using the contrast factor can be used instead of the subjective assessment.