Cardiac O-water PET was performed in Zurich (Discovery ST RX GE Medical SystemMilwaukeekee, WI) with the acquisition of a background frame immediately before the infusion of O-water and a 5-min dynamic acquisition (24 frames, 47 slices each, 2-D mode; 15.2-cm axial field of view; 14 × 5, 3 × 10, 3 × 20, 4 × 30 s).
In the first experiment, we used the ASOD [34] dataset comprised of input frames, a background frame, and the corresponding ground truth (manually annotated and automatically generated inaccurate masks) of static foreground from PETS2006 [35] , PETS2007 [36] , AVSS2007 [27]CVSG [37], VISOR [38], CANDELA [39], and WCAM [40].
The SOBS approach can obtain the modeled background frame of a bootstrapping video sequence with suitable parameter values [27].
Emission data were then acquired over 60 min in list mode as 18 time-frames (one background frame of approximately 7 min prior to injection, followed by 17 frames: 1 × 15 s, 1 × 5 s, 1 × 10 s, 1 × 30 s, 4 × 60 s, 7 × 300 s and 2 × 600 s).
Recently, a dynamic background frame termed as the most common frame of a scene (McFIS) [21] has been developed for video coding using dynamic background modeling.
The RFT requires subjects to set a luminescent rod to the true vertical in the presence and absence of a luminescent background frame.
Starting after the background frame, a dose of 500 700 MBq of 15O-water was injected as an intravenous bolus over 20 s at an infusion rate of 24 mL/min to assess MBF at rest and after bicycle exercise at baseline conditions and after ingestion of caffeine.
However, each modeled background frame B t of the SOBS approach may preserve some foreground objects in I l resulting in some ghost objects in the segmented foreground frame F t.
Figure 9(a) shows the background frame.
(a) Background model of frame 523.
