Accelerating ultrasound image analysis research through publically available database

Ultrasound is widely used intra-operatively to provide real-time feedback in image guided intervention procedures. Registration of preand intra-operative images is a crucial step in the procedure. Unfortunately, real-time US images often have poor signal-to-noise ratio and suffer from imaging artifacts. Hence, registration using US images can be challenging and significant preprocessing is often required to make the registrations robust. The amount of preprocessing required can be reduced by incorporating US physical imaging process. However, progress in this research is hampered due to lack of publicly available database for training and testing image analysis algorithms that take in to consideration ultrasound physical process. We present here a new database that we are building to archive and distribute ultrasound images of an abdominal phantom acquired at different image acquisition parameters. The database contains tracking information of the transducer in addition to the 2D ultrasound image slices. We believe a publicly available database like this one will provide a valuable resource for the research community and it will be instrumental in developing a collaborative scientific community needed to advance the field.