Effect of Blur Setting on Auto-find Performance

High noise levels can severaly hamper auto-find performance, so the blur_lost_and_found setting is used for noisy images. The ability of the auto-find feature to detect spots on a single frame is tested with a grid of 25 Gaussian spots on a 400 by 400 pixel image. The Gaussian spots have a standard deviation 5 and maximum value 0.3, and the background has a brightness of 0.4. Different levels of blurring are tested on three noisy images to observe the effect of the blur setting. The table indicates how many beads out of the possible 25 the Video Spot Tracker detects. For all trials, the Video Spot Tracker maintain_fluorescent_beads option is set at 25. All trials use a symmetric tracking kernel with a radius of 13.

Noise Level	Blur Level	Beads Found	VST Log	Image
Mean=200 Var=1e6	0	25	VST Log	Image
Mean=200 Var=1e6	0.5	25	VST Log	Image
Mean=200 Var=1e6	1	25	VST Log	Image
Mean=200 Var=1e6	2	25	VST Log	Image
Mean=200 Var=1e6	3	25	VST Log	Image
Mean=200 Var=1e6	4	25	VST Log	Image
Mean=200 Var=1e6	5	25	VST Log	Image
Mean=2000 Var=1e7	0	25	VST Log	Image
Mean=2000 Var=1e7	0.5	25	VST Log	Image
Mean=2000 Var=1e7	1	25	VST Log	Image
Mean=2000 Var=1e7	2	25	VST Log	Image
Mean=2000 Var=1e7	3	25	VST Log	Image
Mean=2000 Var=1e7	4	25	VST Log	Image
Mean=2000 Var=1e7	5	25	VST Log	Image
Mean=20000 Var=1e8	0	1	VST Log	Image
Mean=20000 Var=1e8	0.5	1	VST Log	Image
Mean=20000 Var=1e8	1	10	VST Log	Image
Mean=20000 Var=1e8	2	25	VST Log	Image
Mean=20000 Var=1e8	3	25	VST Log	Image
Mean=20000 Var=1e8	4	25	VST Log	Image
Mean=20000 Var=1e8	5	25	VST Log	Image

Example images follow:

gaussian_5_sym_flat.200_1e6.0000.bmp

gaussian_5_sym_flat.2000_1e7.0000.bmp

gaussian_5_sym_flat.20000_1e8.0000.bmp