Fine needle biopsy material segmentation with hough transform and active contouring technique
Maciej Hrebień, Piotr J. Stec · 2006
Automatic diagnostic systems of cancer diseases have a very long history and are still under development by many research and university centres as well as various commercial institutions [3]. Since a breast cancer is becoming the most common women’s disease today, many efforts are still undertaken; to increase its early detecting. The nucleus of the cell is the place where a breast cancer malignancy can be observed. Therefore, it is crucial problem for camera-based automatic diagnostic system; separating the cancer cells from the rest of the image. Until now, many segmentation methods were proposed [2, 6, 7] but unfortunately each of them introduces different kinds of additional problems. Because many cytological projects assume full automation processing and realtime operations implementation, the highest diagnostic level is expected. The methods with deprived drawbacks of already known approaches have to be recognised. The paper discusses a hybrid method that is combining the Hough transformation with the Active Contour method, for cytological image segmentation. The paper also presents some pre-processing steps used in the approach for preliminary experimental procedures collecting some sampling data by the elaborated benchmark database.