New Assessment System for Data Compression in Ultrasound Computer Tomography

This paper proposes a comparison method for various compression techniques for medical imaging systems. The quality of images reconstructed with compressed datasets is used to score the compression process. The image quality is measured by mutual information between a reference and the reconstructed images. Six lossy compression methods, ranging from simple thresholding to discrete wavelet based compression methods were used to compress ultrasound signals over a large range of compression ratios. Experimental results with simulated ultrasound computer tomography system show the influence of compression on datasets is properly measured by mutual information. The possible utilization of the assessment system is discussed.