TY - GEN
T1 - Simultaneous dictionary learning and reconstruction from subsampled data in photoacoustic microscopy
AU - Sathyanarayana, Sushanth G.
AU - Ning, Bo
AU - Hu, Song
AU - Hossack, John A.
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/10
Y1 - 2019/10
N2 - Photoacoustic microscopy acquires volumetric RF data to obtain high resolution, high contrast, images of the microvasculature but is associated with slow acquisition of data due to mechanical raster scanning across the image plane. Recent work has shown that the acquisition speed can be increased using compressive sampling methods and subsequent reconstruction. These methods use bases (dictionaries) learned from prior fully sampled acquisitions, or classical bases such as the Fourier or wavelet bases. In this study, we present the simultaneous learning of bases, and reconstruction using only subsampled data. The algorithm was validated at two different subsampling levels 50% and 75% downsampling, and compared to the ground truth reconstruction with fully sampled data by estimating the peak signal to noise ratio (PSNR). No significant difference in performance was observed between the fully sampled (20.0±3.0 dB), 50% (19.9±2.1 dB) and 75% (19.1±2.6 dB) subsampled data.
AB - Photoacoustic microscopy acquires volumetric RF data to obtain high resolution, high contrast, images of the microvasculature but is associated with slow acquisition of data due to mechanical raster scanning across the image plane. Recent work has shown that the acquisition speed can be increased using compressive sampling methods and subsequent reconstruction. These methods use bases (dictionaries) learned from prior fully sampled acquisitions, or classical bases such as the Fourier or wavelet bases. In this study, we present the simultaneous learning of bases, and reconstruction using only subsampled data. The algorithm was validated at two different subsampling levels 50% and 75% downsampling, and compared to the ground truth reconstruction with fully sampled data by estimating the peak signal to noise ratio (PSNR). No significant difference in performance was observed between the fully sampled (20.0±3.0 dB), 50% (19.9±2.1 dB) and 75% (19.1±2.6 dB) subsampled data.
KW - compressed sensing
KW - Dictionary learning
KW - Photoacoustic Microscopy
UR - https://www.scopus.com/pages/publications/85077624632
U2 - 10.1109/ULTSYM.2019.8925747
DO - 10.1109/ULTSYM.2019.8925747
M3 - Conference contribution
AN - SCOPUS:85077624632
T3 - IEEE International Ultrasonics Symposium, IUS
SP - 483
EP - 486
BT - 2019 IEEE International Ultrasonics Symposium, IUS 2019
PB - IEEE Computer Society
T2 - 2019 IEEE International Ultrasonics Symposium, IUS 2019
Y2 - 6 October 2019 through 9 October 2019
ER -