A safe decision support system based steganographic methods for cells recognition and counting from urine images

Urine analysis is used in the diagnosis of many conditions and diseases of the human body. The manual microscopic urine analysis procedure leads to loss a lot of time together tith the problem of the fallibility of the human eye and therefore this process may result in many erroneous conclusions. Nowadays, using computer algorithms to automate the microscopic urine analysis has gained importance due to these disadvantages. Today, many image processing and classification algorithms are used for the determination and classification of cells existing in the urine sediments. The results of these studies showed that there are some problems associated with urine cells, which reduce the recognition performance, such as adherent cells or overlaping cells. It is require to eliminate these disadvantages in order to have a working algorithm. In this study, many different decision support systems were developed for automatic recognition and counting of cells from real microscobic urine sediment images. These decision support systems has been made more secure by using steganographic methods. They are based on image processing, discrete wavelet transform and intelligent classifiers. Finally, the results obtained from the application of these robust decision support systems were evaluated and compared.