Texture-based features for classification of mammograms using decision tree

Mammogram—breast X-ray—is considered the most effective, low cost, and reliable method in early detection of breast cancer. Although general rules for the differentiation between benign and malignant breast lesions exist, only 15–30 % of masses referred for surgical biopsy are actually malignant. In this work, an approach is proposed to develop a computer-aided classification system for cancer detection from digital mammograms. The proposed system consists of three major steps. The first step is region of interest (ROI) extraction of 256 × 256 pixels size. The second step is the feature extraction; we used a set of 19 GLCM and GLRLM features, and the 19 (nineteen) features extracted from gray-level run-length matrix and gray-level co-occurrence matrix could distinguish malignant masses from benign masses with an accuracy of 96.7 %. Further analysis was carried out by involving only 12 of the 19 features extracted, which consists of 5 features extracted from GLCM matrix and 7 features extracted from GLRL matrix. The 12 selected features are as follows: Energy, Inertia, Entropy, Maxprob, Inverse, SRE, LRE, GLN, RLN, LGRE, HGRE, and SRLGE; ARM with 12 features as prediction can distinguish malignant mass image and benign mass with a level of accuracy of 93.6 %. Further analysis showed that area under the receiver operating curve was 0.995, which means that the accuracy level of classification is good or very good. Based on that data, it was concluded that texture analysis based on GLCM and GLRLM could distinguish malignant image and benign image with considerably good result. The third step is the classification process; we used the technique of decision tree using image content to classify between normal and cancerous masses. The proposed system was shown to have the large potential for cancer detection from digital mammograms.     

A comparison of different Gabor feature extraction approaches for mass classification in mammography

We investigate the performance of six different approaches for directional feature extraction for mass classification problem in digital mammograms. These techniques use a bank of Gabor filters to extract the directional textural features. Directional textural features represent structural properties of masses and normal tissues in mammograms at different orientations and frequencies. Masses and micro-calcifications are two early signs of breast cancer which is a major leading cause of death in women. For the detection of masses, segmentation of mammograms results in regions of interest (ROIs) which not only include masses but suspicious normal tissues as well (which lead to false positives during the discrimination process). The problem is to reduce the false positives by classifying ROIs as masses and normal tissues. In addition, the detected masses are required to be further classified as malignant and benign. The feature extraction approaches are evaluated over the ROIs extracted from MIAS database. Successive Enhancement Learning based weighted Support Vector Machine (SELwSVM) is used to efficiently classify the generated unbalanced datasets. The average accuracy ranges from 68 to 100 % as obtained by different methods used in our paper. Comparisons are carried out based on statistical analysis to make further recommendations.     

A novel neural-genetic algorithm to find the most significant combination of features in digital mammograms

Digital mammography is one of the most suitable methods for early detection of breast cancer. It uses digital mammograms to find suspicious areas containing benign and malignant microcalcifications. However, it is very difficult to distinguish benign and malignant microcalcifications. This is reflected in the high percentage of unnecessary biopsies that are performed and many deaths caused by late detection or misdiagnosis. A computer based feature selection and classification system can provide a second opinion to the radiologists in assessment of microcalcifications. The research in this paper proposes a neural-genetic algorithm for feature selection to classify microcalcification patterns in digital mammograms. It aims to develop a step-wise algorithm to find the best feature set and a suitable neural architecture for microcalcification classification. The obtained results show that the proposed algorithm is able to find an appropriate feature subset, which also produces a high classification rate.     

Mass lesion detection with a fuzzy neural network

This paper presents a novel fuzzy neural network (FNN) approach to detect malignant mass lesions on mammograms. The FNN is a self-adjusting and adaptive system. It is simple in structure and easy to incorporate experts’ knowledge and fuzzified factors in the detection of malignant mass lesions on mammograms. The FNN has four layers. The first layer is the input layer consisting of 4 fuzzy neurons. The second layer has 4 ordinary neurons. The third layer consists of N maximum fuzzy neurons. The number of fuzzy neurons, N, in the third layer is determined during the training process and varies with the network parameters and data distribution. The fourth layer has 2 maximum fuzzy neurons and one competitive fuzzy neuron. Mammograms were obtained from the digital database for screening mammography, DDSM. Six-hundred and seventy regions of interest (ROIs) were extracted from 100 mammograms. All extracted ROIs were randomly divided into two sets: training and testing sets. The co-occurrence matrix of each ROI was computed. Textural features were calculated at sizes of 256×256 and 768×768, respectively. The feature differences at these two image sizes were computed for each feature. These feature differences are very discriminant in differentiating between malignant masses and normal tissues regardless of lesion shape, size, and subtlety. After training, the FNN can correctly detect all malignant masses on mammograms in the testing group. The true-positive fraction (TPF) is 0.92 when the number of false positives (FP) is 1.33 per mammogram and 1.0 when the FP is 2.15 per mammogram. The proposed approach will be very useful for breast cancer control.     

Breast density classification to reduce false positives in CADe systems

This paper describes a novel weighted voting tree classification scheme for breast density classification. Breast parenchymal density is an important risk factor in breast cancer. Moreover, it is known that mammogram interpretation is more difficult when dense tissue is involved. Therefore, automated breast density classification may aid in breast lesion detection and analysis. Several classification methods have been compared and a novel hierarchical classification procedure of combined classifiers with linear discriminant analysis (LDA) is proposed as the best solution to classify the mammograms into the four BIRADS tissue classes. The classification scheme is based on 298 texture features. Statistical analysis to test the normality and homoscedasticity of the data was carried out for feature selection. Thus, only features that are influenced by the tissue type were considered. The novel classification techniques have been incorporated into a CADe system to drive the detection algorithms and tested with 1459 images. The results obtained on the 322 screen-film mammograms (SFM) of the mini-MIAS dataset show that 99.75% of samples were correctly classified. On the 1137 full-field digital mammograms (FFDM) dataset results show 91.58% agreement. The results of the lesion detection algorithms were obtained from modules integrated within the CADe system developed by the authors and show that using breast tissue classification prior to lesion detection leads to an improvement of the detection results. The tools enhance the detectability of lesions and they are able to distinguish their local attenuation without local tissue density constraints.     

Bi-modal breast cancer classification system

Breast cancer is known as one of the major causes of mortality among women. Breast cancer can be treated with better patient outcomes and significantly lower costs if it is detected early. Digital mammograms are the type of medical images most often used, and which are the most reliable, for the detection of breast cancer. The presence of microcalcification clusters in mammograms contributes to evidence for the detection of early stages of cancer. In this paper, a bi-modal artificial neural network (ANN) based breast cancer classification system is proposed. The microcalcifications are extracted with adaptive neural networks that are trained with cancer/malignant and normal/benign breast digital mammograms of both cranio caudal (CC) and medio-latral oblique (MLO) views. The performance of the networks is evaluated using receiver operating characteristic (ROC) curve analysis. Sensitivity–specificity of 98.0–100.0 for the CC view and 96.0–100.0 for the MLO view networks are recorded for 200 unseen digital database for screening mammography (DDSM) cases. The DDSM database, developed at the University of South Florida, is a resource for use by the mammographic image analysis research community. The OR logic is then used to fuse individual networks to get a best sensitivity–specificity of 100.0–100.0 for the ensemble. However, the overall sensitivity–specificity of the ANN ensemble is somewhat degraded at the expense of a robust or sensitive system, i.e., the probability to miss out a true positive case is minimized.     

A statistical framework for breast tumor classification from ultrasonic images

Segmentation and classification of ultrasonic breast images is extremely critical for medical diagnosis. Over the last years, various techniques have already been presented for this objective. In this paper, a proposed framework is presented to segment a given ultrasonic image with breast tumor and classify the tumor as being benign or malignant. The proposed framework depends on an active contour segmentation model to determine the tumor region, and then extract it from the ultrasonic image. After that, the Discrete Wavelet Transform (DWT) is used to extract features from the segmented images. Then, the dimensions of the resulting features are reduced by applying feature reduction approaches, namely, the Principal Component Analysis (PCA), the Linear Discriminant Analysis (LDA) and both of them together. The obtained features are submitted to a statistical classifier and the strategy of voting is used to classify the tumor. In the simulation work, 160 benign and malignant breast tumor images collected from Sirindhorn International Institute of Technology (SIIT) website are used. The average processing time for a 256 × 256 image on a laptop with Core i5, 2.3 GHz processor and 8GB RAM is 1.8 s. From the simulation results, it is found that the utilization of the PCA approach provides the best accuracy of 99.23% among the three feature reduction approaches applied. Finally, the proposed framework is compared with the Support Vector Machine (SVM) classification to evaluate its performance in terms of accuracy, sensitivity, precision, and specificity. It is noticed that the proposed framework is efficient and rapid, and it can be applied for ultrasonic breast image segmentation and classification, and thus it can assist the specialists to segment and decide whether a tumor is benign or malignant.

     

Texture-based analysis of clustered microcalcifications detected on mammograms

The need for early detection of breast cancer has led to establishing screening programs that generate large volumes of mammograms to be analyzed. These analysis are time consuming and labor intensive. Computerized analysis of mammograms has been suggested as “second opinion” or “pre-reader”.In this paper, we suggest a texture-based computerized analysis clusters of microcalcifications detected on mammograms in order to classify them into benign and malignant types.The test of the proposed system yielded a sensitivity of 100%, a specificity of 87.77% and a good classification rate of 89%; the area under the fitted ROC-curve using the MedCalc Statistical Software was 0.968.     

Deep feature extraction and classification of breast ultrasound images

Controlled despeckling (structure/edges/feature preservation with smoothing the homogeneous areas) is a desired pre-processing step for the design of computer-aided diagnostic (CAD) systems using ultrasound images as the presence of speckle noise masks diagnostically important information making interpretation difficult even for experienced radiologist. For efficiently classifying the breast tumors, the conventional CAD system designs use hand-crafted features. However, these features are not robust to the variations in size, shape and orientation of the tumors resulting in lower sensitivity. Thus deep feature extraction and classification of breast ultrasound images have recently gained attention from research community. The deep networks come with an advantage of directly learning the representative features from the images. However, these networks are difficult to train from scratch if the representative training data is small in size. Therefore transfer learning approach for deep feature extraction and classification of medical images has been widely used. In the present work the performance of four pre-trained convolutional neural networks VGG-19, SqueezeNet, ResNet-18 and GoogLeNet has been evaluated for differentiating between benign and malignant tumor types. From the results of the experiments, it is noted that CAD system design using GoogLeNet architecture for deep feature extraction followed by correlation based feature selection and fuzzy feature selection using ANFC-LH yields highest accuracy of 98.0% with individual class accuracy value of 100% and 96% for benign and malignant classes respectively. For differentiating between the breast tumors, the proposed CAD system design can be utilized in routine clinical environment.

     

A novel image mining technique for classification of mammograms using hybrid feature selection

The image mining technique deals with the extraction of implicit knowledge and image with data relationship or other patterns not explicitly stored in the images. It is an extension of data mining to image domain. The main objective of this paper is to apply image mining in the domain such as breast mammograms to classify and detect the cancerous tissue. Mammogram image can be classified into normal, benign, and malignant class. Total of 26 features including histogram intensity features and gray-level co-occurrence matrix features are extracted from mammogram images. A hybrid approach of feature selection is proposed, which approximately reduces 75% of the features, and new decision tree is used for classification. The most interesting one is that branch and bound algorithm that is used for feature selection provides the best optimal features and no where it is applied or used for gray-level co-occurrence matrix feature selection from mammogram. Experiments have been taken for a data set of 300 images taken from MIAS of different types with the aim of improving the accuracy by generating minimum number of rules to cover more patterns. The accuracy obtained by this method is approximately 97.7%, which is highly encouraging.     

Detection of mitotic cells in breast cancer histopathological images using deep versus handcrafted features

One of the most important processes in the diagnosis of breast cancer, which is the leading mortality rate in women, is the detection of the mitosis stage at the cellular level. In literature, many studies have been proposed on the computer-aided diagnosis (CAD) system for detecting mitotic cells in breast cancer histopathological images. In this study, comparative evaluation of conventional and deep learning based feature extraction methods for automatic detection of mitosis in histopathological images are focused. While various handcrafted features are extracted with textural/spatial, statistical and shape-based methods in conventional approach, the convolutional neural network structure proposed on the deep learning approach aims to create an architecture that extracts the features of small cellular structures such as mitotic cells. Mitosis detection/counting is an important process that helps us assess how aggressive or malignant the cancer’s spread is. In the proposed study, approximately 180,000 non-mitotic and 748 mitotic cells are extracted for the evaluations. It is obvious that the classification stage cannot be performed properly due to the imbalanced numbers of mitotic and non-mitotic cells extracted from histopathological images. Hence, the random under-sampling boosting (RUSBoost) method is exploited to overcome this problem. The proposed framework is tested on mitosis detection in breast cancer histopathological images dataset provided from the International Conference on Pattern Recognition (ICPR) 2014 contest. In the results obtained with the deep learning approach, 79.42% recall, 96.78% precision and 86.97% F-measure values are achieved more successfully than handcrafted methods. A client/server-based framework has also been developed as a secondary decision support system for use by pathologists in hospitals. Thus, it is aimed that pathologists will be able to detect mitotic cells in various histopathological images more easily through necessary interfaces.

     

利用支持向量机对癌细胞的识别

癌细胞识别是数字图像处理和模式识别领域的一个研究热点,在临床上也是一个比较有意义的研究课题,本文利用支持向量机的方法,通过提取细胞的特征来分类和识别癌细胞。首先收集彩色图像,对图像进行预处理,之后是进行分割及对单细胞和多细胞的特征提取,最后实现对癌变细胞的识别和分类。     

A Framework of Deep Learning and Selection-Based Breast Cancer Detection from Histopathology Images

Breast cancer (BC) is a most spreading and deadly cancerous malady which is mostly diagnosed in middle-aged women worldwide and effecting beyond a half-million people every year. The BC positive newly diagnosed cases in 2018 reached 2.1 million around the world with a death rate of 11.6% of total cases. Early diagnosis and detection of breast cancer disease with proper treatment may reduce the number of deaths. The gold standard for BC detection is biopsy analysis which needs an expert for correct diagnosis. Manual diagnosis of BC is a complex and challenging task. This work proposed a deep learning-based (DL) solution for the early detection of this deadly disease from histopathology images. To evaluate the robustness of the proposed method a large publically available breast histopathology image database containing a total of 277524 histopathology images is utilized. The proposed automatic diagnosis of BC detection and classification mainly involves three steps. Initially, a DL model is proposed for feature extraction. Secondly, the extracted feature vector (FV) is passed to the proposed novel feature selection (FS) framework for the best FS. Finally, for the classification of BC into invasive ductal carcinoma (IDC) and normal class different machine learning (ML) algorithms are used. Experimental outcomes of the proposed methodology achieved the highest accuracy of 92.7% which shows that the proposed technique can successfully be implemented for BC detection to aid the pathologists in the early and accurate diagnosis of BC.     

Performance Analysis of Breast Cancer Detection Method Using ANFIS Classification Approach

Breast cancer is one of the deadly diseases prevailing in women. Earlier detection and diagnosis might prevent the death rate. Effective diagnosis of breast cancer remains a significant challenge, and early diagnosis is essential to avoid the most severe manifestations of the disease. The existing systems have computational complexity and classification accuracy problems over various breast cancer databases. In order to overcome the above-mentioned issues, this work introduces an efficient classification and segmentation process. Hence, there is a requirement for developing a fully automatic methodology for screening the cancer regions. This paper develops a fully automated method for breast cancer detection and segmentation utilizing Adaptive Neuro Fuzzy Inference System (ANFIS) classification technique. This proposed technique comprises preprocessing, feature extraction, classifications, and segmentation stages. Here, the wavelet-based enhancement method has been employed as the preprocessing method. The texture and statistical features have been extracted from the enhanced image. Then, the ANFIS classification algorithm is used to classify the mammogram image into normal, benign, and malignant cases. Then, morphological processing is performed on malignant mammogram images to segment cancer regions. Performance analysis and comparisons are made with conventional methods. The experimental result proves that the proposed ANFIS algorithm provides better classification performance in terms of higher accuracy than the existing algorithms.     

Computer aided detection system for micro calcifications in digital mammograms

Breast cancer continues to be a significant public health problem in the world. Early detection is the key for improving breast cancer prognosis. Mammogram breast X-ray is considered the most reliable method in early detection of breast cancer. However, it is difficult for radiologists to provide both accurate and uniform evaluation for the enormous mammograms generated in widespread screening. Micro calcification clusters (MCCs) and masses are the two most important signs for the breast cancer, and their automated detection is very valuable for early breast cancer diagnosis. The main objective is to discuss the computer-aided detection system that has been proposed to assist the radiologists in detecting the specific abnormalities and improving the diagnostic accuracy in making the diagnostic decisions by applying techniques splits into three-steps procedure beginning with enhancement by using Histogram equalization (HE) and Morphological Enhancement, followed by segmentation based on Otsu's threshold the region of interest for the identification of micro calcifications and mass lesions, and at last classification stage, which classify between normal and micro calcifications ‘patterns and then classify between benign and malignant micro calcifications. In classification stage; three methods were used, the voting K-Nearest Neighbor classifier (K-NN) with prediction accuracy of 73%, Support Vector Machine classifier (SVM) with prediction accuracy of 83%, and Artificial Neural Network classifier (ANN) with prediction accuracy of 77%.     

基于LVQ神经网络的微钙化分类方法

提出了一个基于自适应的学习矢量量化神经网络（LVQ）的乳腺肿瘤良恶性分类方法,在提取特征向量的基础上,对CC和MLO两种视图的良性和恶性数字化乳腺X光片图像进行训练和测试,并使用最佳分类率和平均分类率来分析分类结果。实验结果表明该方法对CC视图的图像的平均测试分类率为92.6%,而对MLO视图是93.18%。在微钙化分类系统中采用逻辑＂或＂的方式合并两种不同视图下的网络,可以获得的最佳分类性能是94.8%。     

A New Image Mining Approach for Detecting Micro-Calcification in Digital Mammograms

Although mammography is typically the best method to detect breast cancer, it does not recognize 3–20% of the cancer cases. Mammography has established itself as the most efficient technique for detecting tiny cancerous tumor and micro-calcifications are the most difficult to detect since they are very small (0.1–1.0 mm) and they are almost contrasted against the images background. The main purpose of this paper is to provide a new method for the automatic diagnosis of micro-calcification in digital mammograms. It is based on image mining, and the results show 97.35% accuracy, which is improved than the previous works. Tests are based on the standard images data corpus, MIAS. The practical result of this research is registered as an invention in the Patents and Industrial Property Registration Organization numbered as 83119.     

改进Xception模型的乳腺钼靶图像识别研究

乳腺X线摄影技术是早期发现乳腺癌的主要方法,但其结果很大程度上受放射科医师临床诊断经验的限制;基于卷积神经网络对乳腺钼靶图像自动分类的研究可以为放射科医师临床诊断提供意见,然而乳腺癌肿块边缘模糊且良恶性肿块特征差异较小,分类任务面临重重挑战;为了提高乳腺钼靶图像分类的准确率,提出一种基于Xception模型的改进优化算法,改进模型中的残差连接模块,并嵌入Squeeze-and-excitation（SE）注意力机制对模型进行优化;采用优化后的Xception模型并结合迁移学习算法进行乳腺钼靶图像特征提取,并优化全连接层网络进行图像分类,使用公开的乳腺癌图像数据库CBIS-DDSM进行实验,将乳腺钼靶图像自动分为良性和恶性;实验结果表明该方法可以有效提高模型的分类效果,准确率和AUC分别达到了97.46%和99.12%。     

Contrast enhancement and brightness preserving of digital mammograms using fuzzy clipped contrast-limited adaptive histogram equalization algorithm

A novel fuzzy logic and histogram based algorithm called Fuzzy Clipped Contrast-Limited Adaptive Histogram Equalization (FC-CLAHE) algorithm is proposed for enhancing the local contrast of digital mammograms. A digital mammographic image uses a narrow range of gray levels. The contrast of a mammographic image distinguishes its diagnostic features such as masses and micro calcifications from one another with respect to the surrounding breast tissues. Thus, contrast enhancement and brightness preserving of digital mammograms is very important for early detection and further diagnosis of breast cancer. The limitation of existing contrast enhancement and brightness preserving techniques for enhancing digital mammograms is that they limit the amplification of contrast by clipping the histogram at a predefined clip-limit. This clip-limit is crisp and invariant to mammogram data. This causes all the pixels inside the window region of the mammogram to be equally affected. Hence these algorithms are not very suitable for real time diagnosis of breast cancer. In this paper, we propose a fuzzy logic and histogram based clipping algorithm called Fuzzy Clipped Contrast-Limited Adaptive Histogram Equalization (FC-CLAHE) algorithm, which automates the selection of the clip-limit that is relevant to the mammogram and enhances the local contrast of digital mammograms. The fuzzy inference system designed to automate the selection of clip-limit requires a limited number of control parameters. The fuzzy rules are developed to make the clip limit flexible and variant to mammogram data without human intervention. Experiments are conducted using the 322 digital mammograms extracted from MIAS database. The performance of the proposed technique is compared with various histogram equalization methods based on image quality measurement tools such as Contrast Improvement Index (CII), Discrete Entropy (DE), Absolute Mean Brightness Coefficient (AMBC) and Peak Signal-to-Noise Ratio (PSNR). Experimental results show that the proposed FC-CLAHE algorithm produces better results than several state-of-art algorithms.