Ultrasound imaging is an invaluable tool for medical diagnosis and intraoperative guidance. Images are acquired in real time at high frame rates, with no radiation exposure, and at much lower expense than CT and MRI. However, ultrasound imaging suffers from a limited dynamic range: at higher power levels, rigid surgical tools and hyperechogenic tissue in the imaging field get overexposed, while at lower power levels hypoechogenic tissue details are lost.
Similar to ultrasound imaging, optical cameras have a limited dynamic range as well. Imaging shadows requires a long exposure, while imaging highlights requires a short one. High dynamic range (HDR) imaging is a method which extends the dynamic range of optical cameras by combining images taken at multiple exposures. In the last few years, this method has become ubiquitous in smartphones and digital cameras.
In this work, we show how HDR imaging techniques can be applied to ultrasound images to improve their dynamic range, enabling the imaging of media over a wide range of
reflection and attenuation characteristics. We also compare several tone mapping methods, which allow the HDR-US images to be displayed on regular (i.e., low dynamic range) screens, while preserving the visibility of features from across the entire dynamic range. This leads to better visualization for clinicians and can improve the performance of subsequent image processing operations. Therefore, the extended dynamic range of HDR ultrasound (HDR-US) can allow for better diagnosis and procedure guidance.
(a–d) Four of the fifteen ultrasound images acquired at various power levels, and their histograms. In low power settings, the catheter body and the skin surface have good contrast and detail; however, the deeper structures are missing. At the higher power settings, the bone surface and the deeper tissue have good detail, but the catheter and the skin are overexposed and have lost fine detail. (e) Tone-mapped high dynamic range ultrasound image. The deep tissue structures, as well as the catheter and skin details, are easily visible.
Publications:
- A. Degirmenci, D. P. Perrin, and R. D. Howe, “High Dynamic Range Ultrasound Imaging,” International Journal of Computer Assisted Radiology and Surgery, 2018.