A Preprocess Method of External Disturbance Suppression for Carotid Wall Motion Estimation Using Local Phase and Orientation of B-Mode Ultrasound Sequences

Figures

• 

  Figure 1. General workflow of the study.

• 

  Figure 2. B-mode ultrasound image (a) and its local orientation images (b) and local phase images (c, d). Local orientation image with a confidence coefficient (e, f), with a smaller standard deviation (σ1=10 and σ2=102) (c, e), and with a larger standard deviation (σ1=20 and σ2=202) (d, f).

• 

  Figure 3. Flowchart of the synthetic process of the B-mode ultrasound image.

• 

  Figure 4. Clinical B-mode ultrasound image after removing the part of the wall (a) and the synthetic image (b).

• 

  Figure 5. Arterial wall motion (a); external disturbances in the longitudinal direction (b) and radial direction (c); rotation angle (d).

• 

  Figure 6. Several frames of the synthetic B-mode ultrasound sequence (a–e) and their corresponding combining phase images (f–j) with wall motion and disturbances at several time points (0 s, 0.9 s, 1.1 s, 2.2 s, and 3.2 s).

• 

  Figure 7. Comparison of estimated results of the disturbance over one respiratory cycle, for the longitudinal direction (a), radial direction (b), and rotation angle (c). The disturbance obtained by the BM (blue), PM (cyan), SBM (magenta), SPM (yellow), SWBM (green), SWOM (gray), SWPM (orange), and SWPOM (black) methods and the reference (red) are demonstrated.

• 

  Figure 8. Estimated results of wall motion over one respiratory cycle without (a) and with (b) disturbance rejection processing. (b) Comparison between the wall motion estimation obtained from the image sequence without disturbance (cyan) and with disturbance but aligned. The alignments are based on the reference disturbance (red) and the estimated disturbance obtained by the BM (blue), SBM (magenta), SWBM (green) SWOM (gray), SWPM (orange), and SWPOM (black) methods.

• 

  Figure 9. Linear regression line and correlation coefficient R between the reference value of the motion amplitude and the estimation performed by the SWPOM method (a); Bland–Altman plot comparing the motion amplitude estimated by our SWPOM method with the reference (REF) (b).

• 

  Figure 10. Clinical B-mode ultrasound images (a–f) and their corresponding combining phase images (g–l) at several time steps (0, 2, 4, 6, 8, and 10 seconds).

• 

  Figure 11. Extracted disturbances from the clinical image sequence, for the longitudinal direction (a), radial direction (b), and rotation angle (c). The disturbance obtained by the BM (blue), PM (cyan), SBM (magenta), SPM (yellow), SWBM (green) SWOM (gray), SWPM (orange), and SWPOM (black) methods is demonstrated.

• 

  Figure 12. Wall motions estimated from a ten-second clinical B-mode ultrasound image sequence in longitudinal (a) and radial (b) directions after disturbance rejection processing. The disturbance rejections are based on the estimated disturbances demonstrated in Figure 12 which are obtained by the BM (blue), PM (cyan), SBM (magenta), SPM (yellow), SWBM (green), SWOM (gray), SWPM (orange), and SWPOM (black) methods.