Innovative Methods for High-resolution Three-dimensional Ultrasound/Photoacoustic Imaging and Image-guided Intervention
Abstract: Ultrasound (US) imaging and photoacoustic (PA) imaging have been widely applied in diagnostics and tissue evaluations. Three-dimensional US/PA imaging provides a larger field of view (FOV) for comprehensive anatomical information and is receiving more attention recently. Multiple configurations have been proposed to enable 3D US/PA imaging, and various algorithms have been studied to improve image resolution. However, the expensive 2D array transducer or bulky external actuation unit for 3D US/PA imaging devices may not be optimal for hand-held applications. In the popular linear array-based 3D US/PA imaging systems, synthetic aperture focusing (SAF) has been applied to improve resolution in the elevation direction, but the extent of improvement is limited by the size of elevation focal point. US image-guided needle access, with advantages such as low radiation and low cost, is involved in many clinical operations. However, the clinician must be experienced in aligning the needle path and the image plane to ensure a good success rate. To reduce the dependency on user’s experience, there is a need for assistance in needle path and image plane alignment.
My dissertation focuses on developing a hardware configuration based on acoustic reflector and a resolution enhancement algorithm based on deconvolution for 3D US/PA imaging, as well as a mechanism to mechanically align the image path and image plane. On the hardware side, we use a linearly sliding acoustic reflector to redirect ultrasound beam and scan the FOV to enable 3D imaging. On the algorithm side, we adopted the mature Richardson-Lucy deconvolution method to work with simulated point spread function (PSF) to reduce the inconvenience in experimentally measuring PSF. Regarding the alignment-assist mechanism, an acoustic reflector with a slot in the middle is used to redirect the ultrasound beam, allowing the needle passing through the slot to align with the image plane. Experimental data show that the reflector-based 3D imaging hardware is effective in providing 3D images in a compact size, and the deconvolution-based algorithm surpasses the previous state-of-the-art by a large margin. Additionally, the reflector-based alignment assistance offers intuitive and accurate insertion in US image-guided needle access.
Advisor: Dr. Haichong Zhang
Committee: Dr. Loris Fichera, Dr. Iulian Iordachita, Dr. Yihao Zheng
Zoom: https://wpi.zoom.us/j/7187868544?omn=91991529132