Abstract. Cancer is the second leading cause of death in US after cardiovascular disease. Image-based computer-aided diagnosis can assist physicians to efficiently diagnose cancers in early stages. Existing computeraided algorithms use hand-crafted features such as wavelet coefficients, co-occurrence matrix features, and recently, histogram of shearlet coefficients for classification of cancerous tissues and cells in images. These hand-crafted features often lack generalizability since every cancerous tissue and cell has a specific texture, structure, and shape. An alternative approach is to use convolutional neural networks (CNNs) to learn the most appropriate feature abstractions directly from the data and handle the limitations of hand-crafted features. A framework for breast cancer detection and prostate Gleason grading using CNN trained on images along with the magnitude and phase of shearlet coefficients is presented. Particularly, we apply shearlet transform on images and extract the magnitude and phase of shearlet coefficients. Then we feed shearlet features along with the original images to our CNN consisting of multiple layers of convolution, max pooling, and fully connected layers. Our experiments show that using the magnitude and phase of shearlet coefficients as extra information to the network can improve the accuracy of detection and generalize better compared to the state-of-the-art methods that rely on hand-crafted features. This study expands the application of deep neural networks into the field of medical image analysis, which is a difficult domain considering the limited medical data available for such analysis.

This paper presents a method for representation and classification of microscopic tissue images using the shearlet transform. The objective is to automatically process biopsy tissue images and assist pathologists in analyzing carcinoma cells, e.g. differentiating between benign and malignant cells in breast tissues. 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. Based on our approach, each image is represented using the discrete shearlet coefficients and histograms of shearlet coefficients and then used for classification of benign versus malignant tissue images using Support Vector Machines. Our experiments on a publically available database of hystopathological images of human breast shows that our fully automatic approach yields in good classification rates and less complexity compared to other methods.

This paper presents a method to diagnose prostate cancer on multiparametric magnetic resonance imaging (Mp-MRI) using the shearlet transform. The objective is classification of benign and malignant regions on transverse relaxation time weighted (T2W), dynamic contrast enhanced (DCE), and apparent diffusion coefficient (ADC) images. Compared with conventional wavelet filters, shearlet has inherent directional sensitivity, which makes it suitable for characterizing small contours of cancer cells. By applying a multi-scale decomposition, the shearlet transform captures visual information provided by edges detected at different orientations and multiple scales in each region of interest (ROI) of the images. ROIs are represented by histograms of shearlet coefficients (HSC) and then used as features in Support Vector Machines (SVM) to classify ROIs as benign or malignant. Experimental results show that our method can recognize carcinoma in T2W, DCE, and ADC with overall sensitivity of 92%, 100%, and 89%, respectively. Hence, application of shearlet transform may further increase utility of Mp-MRI for prostate cancer diagnosis.

No abstract

The Gleason grading system is generally used for histological grading of prostate cancer. In this paper, we first introduce using the Shearlet transform and its coefficients as texture features for automatic Gleason grading. The Shearlet transform is a mathematical tool defined based on affine systems and can analyze signals at various orientations and scales and detect singularities, such as image edges. These properties make the Shearlet transform more suitable for Gleason grading compared to the other transform-based feature extraction methods, such as Fourier transform, wavelet transform, etc. We also extract color channel histograms and morphological features. These features are the essential building blocks of what pathologists consider when they perform Gleason grading. Then, we use the multiple kernel learning (MKL) algorithm for fusing all three different types of extracted features. We use support vector machines (SVM) equipped with MKL for the classification of prostate slides with different Gleason grades. Using the proposed method, we achieved high classification accuracy in a dataset containing 100 prostate cancer sample images of Gleason Grades 2-5.Gland units contain epithelial cells around the lumen. When cancer happens, epithelial cells randomly duplicate, disturbing the normal structure of glands, as shown in Figure 2b,c.