Skip to main content
. 2019 Mar 20;9:4907. doi: 10.1038/s41598-019-41373-0

Figure 2.

Figure 2

SVD-based clutter filtering of Vevo® 3100 cineloops demonstrated superior microvascular detection and reduced tissue background. (A) B-mode IQ cineloops (100 frames) were considered as three-dimensional imaging volumes with axial, lateral, and slow-time dimensions. (B) These spatiotemporal data sets were reshaped into 2D Casorati matrices in a column-wise manner. Singular value decomposition of Casorati matrices decomposed the data into singular vectors and singular values, allowing for the extraction of blood signals. (C) An inverse SVD was performed to recover blood flow signals back to the spatiotemporal domain, resulting in a superior microvascular image in comparison to conventional power Doppler mode.