New adaptive algorithm for gaussian fitting and color difference in HPTLC plant analysis

Peak overlap is an important cause of erroneous quantification in HPTLC, and color differences of zones are not usually studied. A new adaptive algorithm was developed based on the assumption that the zone due to each component resembles the shape of a gaussian distribution. The chromatogram is the sum of the different gaussian peaks, and the algorithm is able to identify partially overlapped peaks and determine the parameters of every single peak, making a more accurate approximation. Determination of the relative area of each component is done by computing the area of zones. In order to perform substance finding, our algorithm calculates Rfs and uses Delta-E to calculate the color difference of zones. Delta E (ΔE) is a standard measure for color difference, defined by the CIELAB 1976 as the Euclidean distance between two points corresponding to two colors in the L*a*b* space. ΔE color differences below 1 are usually assumed to be not distinguishable by the human vision. The program was implemented in C++ using open-source libraries: CImg library and tinyxml, a free XML parser library, modified by us for supporting stylesheet attaching, since data are presented in a clear format readable by the user with a standard web browser. It was compiled under two platforms: Linux Ubuntu 10.10 GNU GCC compiler version 4.4.3 and Windows 7 Home Edition Microsoft Visual C++ Express 2010.