Preferential color substances and optimized illuminations for computer screen photo-assisted classification

The computer screen photo-assisted technique (CSPT) is a practical method aimed at the evaluation of colorimetric assays using regular computer sets and web cameras as only instruments. The performance of CSPT, understood as its ability to distinguish different color substances, depends of the color of the evaluated substances and on the particular illuminating sequences used for the measurements. In this work, the systematic study of the CSPT performance for sets of 2 and 5 similar color substances is accomplished with the aid of a CSPT model. The result allows identifying specific substance transmittances most favorable with the classification capability of the technique, which is relevant for the design of assays customized for CSPT.The optimization of three colors short illuminating sequences (implying fast measurements) for specific sets of substances is demonstrated and compared with the performance for white illumination. The resulting average improvement in the performance, quantified as the Euclidean distance in a principal components space, for sets of 2 and 5 substances approximately doubles the result with white illumination, and presents performance peaks more than five times larger for critical substance sets.