Figure 2
Performance characterization of the T-PT technique. (a) Analytically prescribed displacement field used for assessing the performance of our T-PT technique. The prescribed displacement field is composed of a sinusoidal field of linearly decreasing spatial wavelength from 200 to 10 pixels. The displacement parameter (d) governs the |u max | of the displacement field. (b) Plot of the recovery ratio of various algorithms versus increasing displacement parameter, d. (c) Plot showing the mismatch ratio of different algorithms against increasing displacement parameter, d. The mismatch ratio for T-PT for d = 0.3 and 0.9 is zero. Since a value of zero is not defined in a log plot, it is indicated by blue stars in the plot. (d) Plot illustrating the execution time of each algorithm against synthetic images seeded with various numbers of particles. The execution time for FVRM and TrackMate for 100,000 particles is not shown in the plot, as the algorithm execution was terminated when its execution time exceeded 105 seconds. (b,c,d) ‘T-PT’ indicates the topology-based particle tracking method introduced here; ‘Legant et al.’ indicates the feature-vector based particle tracking method by Legant et al.;12,15 ‘FVRM’ indicates the feature-vector based relaxation method;34 ‘LAP’ indicates Jaqaman et al.’s LAP-based algorithm24 implemented in TrackMate54. (e) Plot comparing the ability of T-PT and FIDVC51 to recover high spatial frequency motion content. The shaded region indicates the standard deviation of the recovered displacement magnitude. (f) Number of particles required in 2D and 3D images (192 z-slices) to resolve the imposed displacement field at the spatial Nyquist resolution versus camera detector resolution in megapixels.
