Figure 3

A graph of the effects of differences in RQE on fitting precision and fitting bias as measured by fitting simulated data. In the “Uncorrected” approach the RQE values were all set to 1.0 in the fitting. In the “WLS Flat Field” all the pixels had the same gain value, the average of the pixel gain values, the images were flat-field corrected and fitting was done using weighted least squares fitting. In the “RQE Correction” approach the fit included the known RQE and gain values for each pixel. CRLB is the Cramer-Rao lower bound calculated as described in Methods. (A) Fitting precision versus emitter intensity. (B) Magnitude of the bias in fitting versus emitter intensity. (C) Same as (B) but only showing the “WLS Flat Field” and “RQE Correction” results (see legend in B). Error bars are standard deviations estimated from 10 independent simulations for each data point.