Prostate cancer detection and gleason grading of histological images using shearlet transform
Hadi Rezaeilouyeh, Mohammad H. Mahoor, Francisco G. La Rosa, Jun Jason Zhang · 2013
In this paper we propose a method for feature representation and classification of prostate tissue images using the shearlet transform. The objective is to automatically process biopsy tissue images and assist pathologists in analysing carcinoma cells. Furthermore, we will automatically grade the cancer using the Gleason grading system. The main contribution of this paper is the use of a holistic based method instead of going through the tedious task of image segmentation, and the use of the shearlet method instead of traditional signal processing methods. Compared with wavelet filters such as the Gabor filter, shearlet has inherent directional sensitivity which makes it suitable for characterizing small contours of carcinoma cells. By applying a multi-scale decomposition, the shearlet transform captures visual information provided by edges detected at different orientations and multiple scales. Each image is represented using histogram of shearlet coefficients and then used for classification of benign and malignant tissues and also Gleason grading of cancer using support vector machines. Our results show that we can achieve a classification rate of 100% for distinguishing benign from cancer cells and an accuracy of 89% for Gleason grading while we maintain low complexity comparing to other methods.