The multispectral images of human skins, obtained in the Color laboratory of University of Joensuu, were studied in the project. The project offers the preprocessing of the spectral images of skin for their further research using Independent Component Analysis. Retrieving of mixture vectors is important part of Independent Component Analysis. There two methods were implemented in the project, and ICA_analysis_GUI program, realizing these algorithms in real life was developed. Also different spectral images of human skin were analyzed during the work on the project. The analysis spectral images of skin using the ICA method is very interesting and important topic to research. Developing of the problem also gives opportunities for better medical research and diagnostics.