Decomposition-based compression of ultrasound raw-data

Abstract

Sonography techniques use multiple transducer elements for tissue visualization. The signals detected at each element are combined in the process of digital beamforming, requiring that large amounts of data be acquired, transferred and processed. One of the main challenges is reducing the data size while retaining the image contents. For this purpose, we propose a component based model for the raw ultrasonic signals. We show that a decomposition based approach with a suited processing scheme for each component individually, can achieve over twenty-fold reduction of needed data size.