Multifractal analysis emerged as an accurate tool for cancer classification, although the process needs complete automation in order to be applied in medical diagnostics. Once multifractal analysis of metastatic carcinoma is done, results are processed and a neural network classification, based on primary carcinoma, is performed. Neural networks are a proven method for data classification, especially when data is not linearly separable. This paper describes an application of neural networks in order to determine primary cancer in case of intraosseous metastatic cancer, as well as the automation of classification process. Goal of this research is to successfully apply multifractal analysis as an auxiliary diagnostic method.

Keywords: класификација, метастазе, мултифрактална анализа, неуралне мреже
Published on website: 12.6.2014