@ieivanov and I just discussed some desired features for the focus-finding algorithms:
return a normalized "SNR" or "confidence" metric alongside the in-focus slice. We're concerned that low-SNR images may make the focus-finding fail and cause erratic stage motion. I need to think more and test exactly what metric we'll use, but @ieivanov and I agreed that we want the metric to be invariant to (1) the number of pixels, (2) the overall brightness of the sample, and (3) the size of the sample.
a plot=True option to save a small figure of the focus-finding run, likely mid-band power vs. slice, a dot for the peak, and an annotation for the confidence metric.
@ieivanov and I just discussed some desired features for the focus-finding algorithms:
plot=Trueoption to save a small figure of the focus-finding run, likely mid-band power vs. slice, a dot for the peak, and an annotation for the confidence metric.