Fig. 4

Interrupted tracking. The video could be tracked using a single set of settings, but a gap had to be introduced. a Time course of contrast, SD2 of each video frame. No significant drop below the threshold (95%) is observed. b Running standard deviation of the contrast metric (rSD2). It defines how variable is the contrast locally. A significant peak can be seen around t=5 s and after t=8 s. This points to intervals of changing focus, which should be treated carefully. c The resulting deformation and force. A gap is introduced around t=5 s to avoid tracking failure and again after t=8 s. Because the focus reverts to the previous state in the second tracked interval, we can use the same pipette pattern and the same reference frame, thus the forces of both intervals are fully compatible, with identical calibration. d Illustration of BFPtool’s internal treatment of data. On the left, the probe under no load, RBC undeformed (x ₀,Δ x=0). The stiffness of the probe is calculated after the geometrical parameters of the probe (R _c , R₀, R _p ) are measured, yielding \(k=274 \frac {\text {pN}}{{\upmu }\text {m}}\). The RBC acts as a spring in the probe. On the right, the probe at any later time, at 4.2 s and 7.2 s as examples. Note the same pipette pattern is shared by both intervals. The RBC is deformed, the deformation corresponds to the change of distance between the bead centre and the pipette anchor, Δ x(t)=x(t)−x ₀, as illustrated. The force is then calculated as F(t)=k·Δ x(t). Total computational CPU time 152 s. The video file used for this analysis, interrupted.avi, is included in the program repository