| abstract |
An image of a sample (3, Fig. 1) is obtained (S1) from an image sensor (2, Fig. 1), perhaps of a smartphone, having two or more colour channels. Mono-colour arrays corresponding to difference colour channels are extracted (S2), with aggregation of pixels, from the image. In an aspect, a filtered array is determined (S3) as a ratio or difference of two mono-colour images, and the presence or concentration of an analyte in the sample is determined. In another aspect, a mono-colour absorbance array is determined for each colour channel; concentration vectors are determined by multiplying absorbance vectors by a deconvolution matrix, each concentration vector including concentration values corresponding to individual analytes. The deconvolution matrix is determined from calibration samples containing known analyte concentrations, by multiplying a matrix generated from concentration vectors of the calibration samples by the inverse of a matrix generated from absorbance vectors based on measured absorbances. |