Figure 2.
((a), left) Evaluation of the unfiltered lighting bias both horizontally and vertically across DI water control μPAD channels within a 12 cm × 12 cm viewing region. Each channel signal intensity (Is) and the three reference intensities (Ib, Ip, and Iw) used within our mathematical normalization scheme were measured for each channel, and then were fractionally compared with the same measurement from the focal center channel. The maximum quantified error by normalization without initial image filtering was 12% ((a), right) To mitigate this lighting bias, a fast-Fourier transform (FFT) bandpass filter was applied to the original image, which reduced the error between channels and the focal center to under ± 1%. (b) This FFT bias correction alone allowed for a more uniform intensity signal measurements both between adjacent μPAD channels and (c–e) vertically within each single channel. This enabled the use of triple-reference point normalization within each channel as well as comparison between channels.