Developing a subsystem for dermatological diseases diagnosis using artificial neural networks

Nowadays, huge distress in dermatology field is caused by diagnose of skin diseases. Due to similarities of skin diseases, wrong diagnoses are caused. Naturally this condition brings wrong treatment. In this study, tried to construct a system about skin diseases recognition. In this working, human skin image are used as a material. To reach attributes from these skin images, attribute matrix from each image using Gabor wavelet lug. Standard deviation, mean and entropi values of attribute matrix. With these values, made a new attribute vector that has 3 parameters for each image. To classification of obtained attributes, used artificial neural networks. Obtained attributes are subjected cross validate test with different learning algorithms and finally reached approximately %85 success rate. Key Words: Artificial Neural Network, Skin Diseases, Gabor Wavelet Lug, Multilayer Perceptron, Cross Validate