The connection between certain human diseases and abnormal changes in chromosomes was discovered nearly 50 years ago. However, the detection of these abnormalities turned out to be a difficult task because both microscopic technologies and biochemical diagnostic techniques have their limitations. In the beginning of the century, a method for multiple comparative genome hybridization (array CGH) was introduced and is now been widely used in clinical practice for biochemical diagnostics. The application of array CGH greatly reduces the statistical errors, but not the noise in the results. As such, the time and material expenses necessary for reliable localization of abnormal parts of chromosomes by the means of biochemical techniques, is not significantly reduced. Algorithms for localization of contrast parts in noisy sequences can improve the situation. We implemented and tested two algorithms for this purpose. The first is the "edge detector" algorithm introduced by J. Canny. The second one is the DotHelix algorithm developed by A. M. Leontovich, L. I. Brodsky, and A. E. Gorbalenya. In this paper we compare the two implementations against biochemical method in processing data produced by array CGH.