Diagnosing system of melanocytic skin lesions

       This system employs user interface in the form of a website to get the access to its three main working layers. The first layer is dedicated to persons without medical background, and serves to self-diagnosing. This layer consist of two modules: the first one allows to determine – in a very simple and clear way – all symptoms required for correct classification of a given skin lesion. Thus, using this module, user can easily acquainted with the knowledge, required for correct assignment of symptoms (asymmetry, border, colours, and diversity of structures), related to a given lesion.
     The second module plays a role of an advanced calculator for noninvasive diagnosing of melanocytic lesions. Input for this module create a vector conveying values of 13 descriptive attributes (asymmetry, border, 6 colour, 5 structures). These values, inputted by the user, are processed tocalculate the 14-th attribute, the TDS (in a fact, also new TDS). Then five different algorithms mentioned previously, are responsible for development of five various learning models (five partial classificators). The classification process based on these models is described in paragraph 3. The second layer, dedicated to dermatology specialists, contains only the second module described above.
       The third layer enables to synthesize images of skin lesion images, by the semantic conversion of records from the textual informational database into respective digital images. In the next stage of our research the fourth layer is planned. This layer will be based on automatic analysis and recognition of melanocytic images.

(use Internet Explorer)