Twin support vector machines and subspace learning methods for microcalcification clusters detection

Microcalcification cluster detection using multiscale products based Hessian matrix via the Tsallis thresholding scheme

Clustered microcalcifications are often the first sign of breast cancer in a mammogram. Nevertheless, all clustered microcalcifications are not found by an individual radiologist reading a mammogram. The use of a second reader may find those clusters of microcalcifications not found by the first reader, thereby improving the sensitivity of detecting clustered microcalcifications. Our laboratory has developed a computerized scheme for the detection of clustered microcalcifications, which can act like a second reader, that is undergoing clinical evaluation. This paper concerns the feature analysis stage of the computer scheme, which is designed to remove some of the false-computer detections. We have examined three methods of feature analysis, namely, rule based (the method currently used), an artificial neural network (ANN), and a combined method. In an independent database of 50 images, at a sensitivity of 83%, the average number of false positive (FP) detections per image was: 1.9 for rule-based, 1.6 for ANN, and 0.8 for the combined method. We demonstrate that the combined method performs best because each of the two stages eliminates different types of false positives.

abstractUniversum, a third class not belonging to either class of the classification problem, allows to incorporate the prior knowledge into the learning process. A lot of previous work have demonstrated that the Universum is helpful to the supervised and semi-supervised classification. Moreover, Universum has already been introduced into the support vector machine (SVM) and twin support vector machine (TSVM) to enhance the generalisation performance. To further increase the generalisation performance, we propose a least squares TSVM with Universum data (-TSVM) in this paper. Our -TSVM possesses the following advantages: first, it exploits Universum data to improve generalisation performance. Besides, it implements the structural risk minimisation principle by adding a regularisation to the objective function. Finally, it costs less computing time by solving two small-sized systems of linear equations instead of a single larger-sized quadratic programming problem. To verify the validity of our proposed algorithm, we conduct various experiments around the size of labelled samples and the number of Universum data on data-sets including seven benchmark data-sets, Toy data, MNIST and Face images. Empirical experiments indicate that Universum contributes to making prediction accuracy improved even stable. Especially when fewer labelled samples given, -TSVM is far superior to the improved LS-TSVM (ILS-TSVM), and slightly superior to the -TSVM.

Detection of clustered microcalcifications in small field digital mammography

Early detection of breast cancer, a significant public health problem in the world, is the key for improving breast cancer early prognosis. Mammography is considered the most reliable and widely used diagnostic technique for early detection of breast cancer. However, it is difficult for radiologists to perform both accurate and uniform evaluation for the enormous mammograms with widespread screening. Microcalcification clusters is one of the most important clue of the breast cancer, and their automated detection is very helpful for early breast cancer diagnosis. Because of the poor quality of the mammographic images and the small size of the microcalcifications, it is a very difficult task to perform detecting the early breast cancer. In this paper, we propose a novel approach based on hybrid subspace fusion for detection microcalcification clusters, and successfully apply it to detection task in digital mammograms. In such a system, subspace learning algorithms will be selectively fused according to the ability of preserving the classification information. Experimental results show that the proposed method improved the performance and stability of microcalcification cluster detection and could be adapt to the noise environments better. The proposed methods could get satisfactory results on sensitivity and reduce false positive rate, which provide some new ideas and methods for the research and development of computer-aided detection system in the breast cancer detection community.

The introduction of multiscale techniques to signal and image processing has provided a new tool to create innovative methods for solving problems in the areas of data compression, signal analysis, and noise removal. Although these techniques are popular and used extensively in research and in engineering applications, their use in signature detection and classification is still an area open to extensive investigation. This paper discusses multiscale techniques working in synergy with other processing techniques to detect and recognize abnormal and cueing signatures that are important to diagnostic medicine-detection and recognition of microcalcification clusters in mammograms. In this application, an innovative detection algorithm that takes advantage of multiresolution analysis and synthesis is developed to assist radiologists looking for clusters of microcalcifications in digitized mammograms. The algorithm presented in this paper successfully limits the false positives. An algorithm description and examples are shown in this paper.

Twin support vector machine (TWSVM) is a classical distance metric learning method for classification problems. The formulation of TWSVM criterion is based on L2-norm distance, which makes TWSVM prone to being influenced by the presence of outliers. In this paper, to develop a robust distance metric learning method, we propose a new objective for TWSVM classifier using L1-norm distance metric, termed as L1-TWSVM. The optimization strategy is to maximize the ratio of the inter-class distance dispersion to the intra-class distance dispersion by using L1-norm distance rather than L2-norm distance. Besides, we design a simple and valid iterative algorithm to solve L1-norm optimal problems, which is easy to actualize and its convergence to an optimum is theoretically ensured. The efficiency and robustness of L1-TWSVM have been validated by experiments on UCI datasets and artificial datasets. The promising experimental results indicate that our proposals outperform relevant state-of-the-art methods in all kinds of experimental settings.

Linear structures are a major contributor to false-positives (FPs) in detection of clustered microcalcifications (MCs) in mammograms. We propose a unified classifier approach to incorporate the dichotomous effect of linear structures in MC detection, the purpose being to suppress the FPs associated with linear structures. We introduce a dummy variable in the classifier model as in traditional regression analysis, the role of which is to adapt the input features to the classifier according to the presence of linear structures. In the experiment we demonstrate the proposed approach by using Adaboost decision stumps as the unified classifier. The results on a set of 200 mammogram images (all containing clustered MCs) show that it could reduce the FPs in an existing SVM detector by up to 47.3% with the true-positive rate at 85%.

Twin support vector machine (TSVM) has been widely applied to classification problems. But TSVM is sensitive to outliers and is not efficient enough to realize feature selection. To overcome the shortcomings of TSVM, we propose a novel sparse twin support vector machine with the correntropy-induced loss (C-STSVM), which is inspired by the robustness of the correntropy-induced loss and the sparsity of the \(\ell _1\)-norm regularization. The objective function of C-STSVM includes the correntropy-induced loss that replaces the hinge loss, and the \(\ell _1\)-norm regularization that can make the decision model sparse to realize feature selection. Experiments on real-world datasets with label noise and noise features demonstrate the effectiveness of C-STSVM in classification accuracy and confirm the above conclusion further.

At present, mammography is the only not invasive diagnostic technique allowing the diagnosis of a breast cancer at a very early stage. A visual clue of such disease particularly significant is the presence of clusters of microcalcifications. Reliable methods for an automatic detection of such clusters are very difficult to accomplish because of the small size of the microcalcifications and of the poor quality of the digital mammograms. A method designed for this task is described. The mammograms are firstly segmented by means of the tree structured Markov random field algorithm which extracts the elementary homogeneous regions of interest on the image. Such regions are then submitted to a further analysis (based both on heuristic rules and support vector classification) in order to reduce the false positives. The approach has been successfully tested on a standard database of 40 mammographic images, publicly available.

Microcalcification (MC) clusters in mammograms can be important early signs of breast cancer in women. Accurate detection of MC clusters is an important but challenging problem. In this paper, we propose the use of a recently developed machine learning technique -- relevance vector machine (RVM) -- for automatic detection of MCs in digitized mammograms. RVM is based on Bayesian estimation theory, and as a feature it can yield a decision function that depends on only a very small number of so-called <i>relevance vectors</i>. We formulate MC detection as a supervised-learning problem, and use RVM to classify if an MC object is present or not at each location in a mammogram image. MC clusters are then identified by grouping the detected MC objects. The proposed method is tested using a database of 141 clinical mammograms, and compared with a support vector machine (SVM) classifier which we developed previously. The detection performance is evaluated using the free-response receiver operating characteristic (FROC) curves. It is demonstrated that the RVM classifier matches closely with the SVM classifier in detection performance, and does so with a much sparser kernel representation than the SVM classifier. Consequently, the RVM classifier greatly reduces the computational complexity, making it more suitable for real-time processing of MC clusters in mammograms.

The condition monitoring and fault diagnosis of power transformers plays a significant role in the safe, stable and reliable operation of the whole power system. Dissolved gas analysis (DGA) methods are widely used for fault diagnosis, however, their accuracy is limited by the selection of DGA features and the performance of fault diagnosis models, for example, the classical support vector machine (SVM), is easily affected by unbalanced training samples. This paper presents a transformer fault diagnosis model based on chemical reaction optimization and a twin support vector machine. Twin support vector machines (TWSVMs) are used as classifiers for solving problems involving unbalanced and insufficient samples. Restricted Boltzmann machines (RBMs) are used for data preprocessing to ensure the effective identification of feature parameters and improve the efficiency and accuracy of fault diagnosis. The chemical reaction optimization (CRO) algorithm is used to optimize TWSVM parameters to select the optimal training parameters. The cross-validation (CV) method is used to ensure the reliability and generalization ability of the diagnostic model. Finally, the validity of the model is verified using real fault samples and random testing.

Multiple instance learning algorithms have been increasingly utilized in many applications. In this paper, we propose a novel multiple instance learning method called GM-Citation-KNN for the microcalcification clusters (MCCs) detection and classification in breast images. After image preprocessing and candidates generation, features are extracted from the potential candidates based on a constructed graph. Then an improved version of Citation-KNN algorithm is used for classification. Regarding each bag as a graph, GM-Citation-KNN calculate the graph similarity to replace the Hausdoff distance in Citation-KNN. The graph similarity is computed by many-to-many graph matching which allows the comparison of parts between graphs. The proposed algorithms were validated on the public breast dataset. Experimental results show that our algorithm can achieve a superior performance compared with some state-of-art MIL algorithms.

We propose computer-aided detection (CADe) algorithm for microcalcification (MC) clusters in reconstructed digital breast tomosynthesis (DBT) images. The algorithm consists of prescreening, MC detection, clustering, and false-positive (FP) reduction steps. The DBT images containing the MC-like objects were enhanced by a multiscale Hessian-based three-dimensional (3D) objectness response function and a connected-component segmentation method was applied to extract the cluster seed objects as potential clustering centers of MCs. Secondly, a signal-to-noise ratio (SNR) enhanced image was also generated to detect the individual MC candidates and prescreen the MC-like objects. Each cluster seed candidate was prescreened by counting neighboring individual MC candidates nearby the cluster seed object according to several microcalcification clustering criteria. As a second step, we introduced bounding boxes for the accepted seed candidate, clustered all the overlapping cubes, and examined. After the FP reduction step, the average number of FPs per case was estimated to be 2.47 per DBT volume with a sensitivity of 83.3%.

KNN weighted reduced universum twin SVM for class imbalance learning