Search

Melanoma is one of the deadliest cancer to be treated as well to detect in initial phase. Here we take the skin lesion by ROI and then we take out features of it then it needs to be segmented whether the particular image is cancerous or not. If it is cancerous then classify the extracted features and discuss about type of stages. A novel 3D reconstruction algorithm from 2D dermoscopic images is proposed. The detection of 3D image shape and RGB are to be carried out. In this paper we have proposed this work for 3D depth parameter which will enhance the classification rate.

This paper is aimed at the development of an approach of applying the curvelet transform to images of prostatectomy pathological specimens of critical Gleason grades for computer-aided classification. A set of Tissue MicroArray (TMA) images from the Johns Hopkins University have been used as the data base. We utilize a moving window to sample multiple patches of a given image leading to a majority decision by the patches for image class assignment. The curvelet-based feature extraction may capture both textural and, implicitly, structural information in an image patch. A tree-structured classifier consisting of three Gaussian-kernel support vector machines each with an embedded voting mechanism has been successfully trained and tested yielding high accuracy to classify tissue images of four critical Gleason scores (GS) 3+3, 3+4, 4+3 and 4+4. The experimental result has demonstrated an enhanced performance as compared to other reported works.

Breast cancer, the most common type of cancer in women is one of the leading causes of cancer deaths. Due to this, early detection of cancer is the major concern for cancer treatment. The most common screening test called mammography is useful for early detection of cancer. It has been proven that there is potential raise in the cancers detected due to consecutive reading of mammograms. But this approach is not monetarily viable. Therefore there is a significant need of computer aided detection systems which can produce intended results and assist medical staff for accurate diagnosis. In this research we made an attempt to build classification system for mammograms using association rule mining based on texture features. The proposed system uses most relevant GLCM based texture features of mammograms. New method is proposed to form associations among different texture features by judging the importance of different features. Resultant associations can be used for classification of mammograms. Experiments are carried out using MIAS Image Database. The performance of the proposed method is compared with standard Apriori algorithm. It is found that performance of proposed method is better due to reduction in multiple times scanning of database which results in less computation time. We also investigated the use of association rules in the field of medical image analysis for the problem of mammogram classification.

Automated prediction of patient-specific disease progression can significantly contribute to clinical treatment. This paper presents a computer-assisted framework to tackle the survival time prediction problem. Inspired by the assumption that niche tumor regions may play a significant role in cancer diagnosis, we explore local visual variations from multiple MRI sequences. The research consists of three parts: 1) the extraction of multi-scale Local Binary Patterns (LBP) to describe the visual variations, 2) a supervised forward feature selection approach, called the Feature Ranking Model (FRM) which captures single feature predictive ability efficiently, and combines the top features to form a feature subset, 3) We cast the clinical survival time prediction task as a binary category classification problem. We tested the framework using a dataset of 32 cases collected from The Cancer Genome Atlas (TCGA). We obtained a 93.75% accuracy rate for the prediction of survival time.

Lymph node is one of the important organs as the immunologic filter in the body. The pathological change of lymph node is an important basis of malignant tumor detection and judgment of metastasis of cancer (lung cancer, colorectal cancer, breast cancer, liver cancer, cervical cancer, etc.). Therefore, the feature analysis of lymph node is meaningful for the forecast about tumor recrudescence and metastasis. Based on image features, lymph node characters and advice from doctors, the sixteen features including shape, texture and grey are extracted and analyzed using cloud model. Eight helpful features (the largest area ratio, length to diameter ratio, the area ratio of length to diameter, the standard deviation of the internal texture changes, standard differential gray, gray consistency, low density values, high density value) for classification of benign and malignant lymph nodes are filtered to recognize pelvic lymph nodes. Improved FCM clustering algorithm is used to detect classification results of eight features achieved by cloud model analysis. Experiment results show that the clustering classification ability of eight features selected by cloud model can achieve an average recognition rate of 70% which is better than that of other features. So the eight features can be used for computer-aided classification and recognition system of pelvic lymph nodes.

Early detection of breast cancer is carried out by using mammographic images. Due to low contrast nature of these images, it is difficult to detect signs such as micro calcifications and masses. This paper describes novel algorithms for early detection of breast cancer using image processing techniques. Novel algorithms are implemented for 1) Mass region extraction to get exact shape of the mass 2) Superposition of boundary of mass on mammogram helps doctors to view the boundary easily as mass region overlaps with breast parenchyma 3) Extraction of texture features like mean, standard deviation, entropy, kurtosis etc, geometric features like area perimeter L:S, ENC, (Elliptical normalized circumference) wavelet based features, so that signatures can be assigned for identification and classification of benign and malignant masses. Fourteen patients' mammograms have been processed. Features of six patients have been extracted that have masses.

The malignant tumors are complex structures, which evolve chaotically, invading the entire human body. The gold standard for cancer diagnosis is the biopsy, but this is invasive, dangerous. We elaborated non-invasive, computerized methods, for tumor characterization, based on ultrasound images. We defined the textural model of the malignant tumors, consisting of the relevant textural features, able to distinguish these structures from similar tissues, and of the specific values associated to the relevant features [1]. In this paper, we analyzed the role that some multiresolution textural features have in improving the liver tumors' diagnosis accuracy. In the new attribute set we added features derived from the second and superior order GLCM and edge-based statistics, all computed after applying the Wavelet transform. The experiments were performed on ultrasound images of patients suffering from hepatocellular carcinoma and from benign liver tumors, considering also the aspect of the cirrhotic parenchyma where the tumors evolve.

This paper presents an algorithm for classification of non-melanoma skin lesions based on a novel hierarchical K-Nearest Neighbors (K-NN) classifier. The K-NN classifier is simple, quick and effective. The hierarchical structure decomposes the classification task into a set of simpler problems, one at each node of the classification. Feature selection is embedded in the hierarchical framework that chooses the most relevant feature subsets at each node of the hierarchy. Colour and texture features are extracted from skin lesions. The accuracy of the proposed hierarchical scheme is higher than 93% in discriminating cancer and pre-malignant lesions from benign lesions, and it reaches an overall classification accuracy of 74% over five common classes of skin lesions, including two non-melanoma cancer types. This is the most extensive published result on non-melanoma skin cancer classification from colour images acquired by a standard camera (non-dermoscopy).

Features extracted from cell networks have become popular tools in histological image analysis. However, existing features do not take sufficient advantage of the cycle structure present within the cell networks. We introduce a new class of network cycle features that take advantage of such structures. We demonstrate the utility of these features for automated prostate cancer scoring using histological images. Prostate cancer is commonly scored by pathologists using the Gleason grading system and our automated system based upon network cycle features serves an important need in making this process less labor-intensive and more reproducible. Our system first extracts the cells from the histological images, computes networks from the cell locations and then computes features based upon statistics for the different cycles present in these networks. Using an SVM (Support Vector Machine) classifier on these features, we demonstrate the efficacy of our system in distinguishing between grade 3 and grade 4 prostate tumors. We also show the superiority of our approach over previously developed systems for this problem based upon texture features, fractal features and alternative network features.

This study presents a computer-aided detection (CADe) system of hepatocellular carcinoma (HCC) using sequential forward floating selection (SFFS) method with linear discriminant analysis (LDA). We extracted morphologic and texture features from the segmented HCC candidate regions from the arterial phase (AP) images of the contrast-enhanced hepatic CT images. To select the most discriminatory features for classification, we developed an SFFS method directly coupled with LDA that maximizes the area under the receiver-operating-characteristic curve (AUC) value. The maximal AUC value criterion directly reflects the CADe system performance used in clinical practice. The initial CADe before the classification achieved a 100% (23/23) sensitivity with 33.7 (775/23) false positives (FPs) per patient. The maximal AUC SFFS method for LDA with eleven selected features eliminated 48.0% (372/775) of the FPs without any removal of the HCCs in a leave-one-lesion-out cross-validation test; thus, a 95.6% sensitivity with 7.9 FPs per patient was achieved.

Cervical cancer is one of the deadliest cancer known and is also a key research area in image processing. The main problem with this cancer is that it cannot be detected as it doesn't throw any symptoms until the final stages. This is attributed to the cancer itself and also to the lack of pathologists available to screen the cancer. Here we have proposed a novel approach to classify the various malignancies in cervical cyto images using the textural properties of the cervical cyto image. For grouping the stages of the cancer we have employed a decision based support system that would help classify the stages of the cancer and help the pathologist detect the cancer better. The proposed image has been tested with a set of images and has proved to be efficient.

Computer vision-based diagnosis systems have been widely used in dermatology, aiming at the early detection of skin cancer and more specifically the recognition of malignant melanoma tumor. This paper proposes a novel clustering technique for the characterization and categorization of pigmented skin lesions in dermatological images. Appropriate image processing techniques (i.e. segmentation, border detection, color and texture processing) are utilized for feature extraction. The proposed method uses Principal Component Analysis and is considered appropriate, since it is suitable for problems with high dimensional data. Initial experimental results have proved the superiority of this method against traditional ones.

Breast cancer is the most frequently diagnosed cancer and the most common cause of cancer death among women. In developing countries, testing for detection of this cancer involves visual microscopic test of cytology samples such as Fine Needle Aspiration Cytology (FNAC) taken from the patient's breast. The result of analysis on this sample by Cyto-pathologist is crucial for breast cancer patient. The wavelet transform has become an important tool for many biomedical image and signal application, which at the detection and analysis of image features. In this paper, clusters of nucleus in the sub-band images of FNAC samples is investigated after decomposition by means of the Complex Discrete Wavelet Transform (CDWT) and a new method is developed to compute the variability and other nucleus statistical textural features. The experimental data are taken from FNAC samples of benign and malignant cases. The calculated variance parameters and texture features permit the characterization of nucleus clusters variable and distinguish malignant samples from benign in cytological breast fine needle aspirate images.

In this paper, a novel automated method to recognize centroblast (CB) cells from non-centroblast (non-CB) cells in follicular lymphoma cases is developed and its performance is evaluated against consensus of 30 board-certified hematopathologists. Morphometric and color texture features are used in the training and testing of a supervised quadratic discriminate analysis (QDA) classifier. The novelty of our method resides in the identification of the CB cells with prior information, and the introduction of the principal component analysis (PCA) in the spectral domain to extract texture color features. A graphical user interface was developed to display CB and non-CB cells without the computer-classification to the hematopathologists and their responses were recorded by the software. Our automated grading system performed well when compared to consensus diagnosis of 30 hematopathologists. Automated classification can identify centroblast cells (CB) from non-centroblast cells (non-CB) with a sensitivity and specificity of 81.8%, 86.4%, respectively. The developed system was tested on an independent set of cases with a consensus of 16 or 20 hematopathologists. The sensitivity and specificity of the developed system is higher when the ground truth is based on the consensus of 20 pathologists.

The objective of this study is to investigate the use of pattern classification methods for distinguishing different types of brain tumors, such as primary gliomas from metastases, and also for grading of gliomas. A computer-assisted classification method combining conventional magnetic resonance imaging (MRI) and perfusion MRI is developed and used for differential diagnosis. The proposed scheme consists of several steps including ROI definition, feature extraction, feature selection and classification. The extracted features include tumor shape and intensity characteristics as well as rotation invariant texture features. Features subset selection is performed using Support Vector machines (SVMs) with recursive feature elimination. The binary SVM classification accuracy, sensitivity, and specificity, assessed by leave-one-out cross-validation on 102 brain tumors, are respectively 87%, 89%, and 79% for discrimination of metastases from gliomas, and 87%, 83%, and 96% for discrimination of high grade from low grade neoplasms. Multi-class classification is also performed via a one-versus-all voting scheme.

Most of the thyroid nodules are heterogeneous with various internal components, which confuse many radiologists and physicians with their various echo patterns in thyroid nodules. A lot of texture extraction methods were used to characterize the thyroid nodules. Accordingly, the thyroid nodules could be classified by the corresponding textural features. In this paper, five support vector machines (SVM) were adopted to select the significant textural features and to classify the nodular lesions of thyroid. Experimental results showed the proposed method classifies the thyroid nodules correctly and efficiently. The comparison results demonstrated that the capability of feature selection of the proposed method was similar to the sequential floating forward selection (SFFS) method. However, the proposed method is faster than the SFFS method.

In this paper, we propose a new computer aided diagnosis method for prostate cancer detection in ultrasound image. With multi resolution autocorrelation texture features and clinical features such as location and shape of tumor, we could maintain high specificity with high sensitivity for prostate cancer detection. Multi resolution autocorrelation can detect cancer suspicious region efficiently with high specificity and sensitivity. And clinical features filters out false positive region by prior knowledge of location and the shape of prostate cancer. Those features are put to Support Vector Machine to classify cancer region or non-cancer region. The proposed method will be helpful in making a more reliable diagnosis,and increasing diagnosis efficiency.

Microcalcification on X-ray mammogram is a significant mark for early detection of breast cancer. Texture analysis methods can be applied to detect clustered microcalcification in digitized mammograms. In order to improve the predictive accuracy of the classifier, the original number of feature set is reduced into smaller set using feature reduction techniques. In this paper rough set based reduction algorithms such as , Quickreduct (QR) and proposes Modified Quickreduct (MQR) are used to reduce the extracted features. The performance of both algorithms is compared. The Gray Level Co-occurrence Matrix (GLCM) is generated for each mammogram to extract the Haralick features as feature set. The reduction algorithms are tested on 161 pairs of digitized mammograms from Mammography Image Analysis Society (MIAS) database.

While microcalcifications (MCs) are important early signs of breast cancers, their reliable detection from mammograms has been largely elusive for both radiologists and computer-aided diagnosis (CAD) strategies. Two of the essential components in a CAD system are the detection of the suspicious MC pixels/regions using image processing and analysis techniques, and the training, classification, and recognition of these areas based on pattern recognition methods. In this paper, we present a novel scheme to identify and classify microcalcifications based on localized texture comparison. Relying on a texture removal and repairing (R&R) process of the preselected suspicious areas from their surrounding background tissues, pre- and post- R&R local characteristic features of these areas are extracted and compared. A modified AdaBoost algorithm is then adopted to train the classifier using expert-labelled microcalcifications, followed by a clustering process. Experiments with the mammographic images from the MIAS and DDSM databases have shown very promising results.

The fluorescence confocal microendoscope (CM) is a new type of instrument for imaging the surface of the human ovary and has diagnostic implications for the early detection of ovarian cancer. The purpose of this study was to develop an automated system to facilitate the identification of ovarian cancer from digital images captured with the CM system. We modeled the cellular distribution present in the images as texture and extracted features based on first-order statistics, spatial gray-level dependence matrices, and spatial-frequency content. We believe this is the first time that automated texture analysis has been used to detect ovarian cancer in CM images. Experiments were conducted to select the best features for classification and to compare the performance of machine classifiers. The results show that it is possible to automatically identify ovarian cancer based on texture features extracted from CM images and that the machine performance is superior to that of the human observer.

This paper presents a comparison of feature selection methods for a unified detection of breast cancers in mammograms. A set of features, including curvilinear features, texture features, Gabor features, and multi-resolution features, were extracted from a region of 512×512 pixels containing normal tissue or breast cancer. Adaptive floating search and genetic algorithm were used for the feature selection, and a linear discriminant analysis (LDA) was used for the classification of cancer regions from normal regions. The performance is evaluated using A z , the area under ROC curve. On a dataset consisting 296 normal regions and 164 cancer regions (53 masses, 56 spiculated lesions, and 55 calcifications), adaptive floating search achieved A z =0.96 with comparison to A z =0.93 of CHC genetic algorithm and A z =0.90 of simple genetic algorithm.

We propose an approach using artificial neural networks to classify masses in mammograms as malignant or benign. Single-layer and multi-layer perceptron networks were used in a study on perceptron topologies for pattern classification of breast masses. The boundaries of 108 breast masses and tumors were manually delineated and represented by polygonal models for shape analysis. Ribbons of pixels were extracted around the boundary of each mass. Three shape factor measures based on the contours, and fourteen texture features based on gray-level co-occurrence matrices of the pixels in the ribbons were computed. Various combinations of the features were used with perceptrons of several topologies for classification of benign masses and malignant tumors. The results were compared in terms of the area A/sub Z/ under the receiver operating characteristics curve. Values of A/sub Z/ up to 0.99 were obtained with the shape factors, whereas texture features provided A/sub z/ up to only 0.63.

Breast cancer is a major cause of cancer deaths among women. Early detection of the primary tumor is an essential and effective method to reduce mortality. Here, the authors present a new automated method for detection of tumors in digital mammograms based on the application of multichannel filtering for texture feature extraction. The channel filters are represented by a computationally efficient infinite impulse response (IIR) QMF bank. The texture feature extraction method is applied to detect stellate lesions in mammograms from the MIAS database. The experiments demonstrate that the authors' approach can provide a true detection rate of approximately 86% and 0 false detections per image for fatty-glandular mammograms.

Breast cancer is the leading cause of death among women. Mammography is the only effective and viable technique to detect breast cancer, sometimes before the cancer becomes invasive. About 30% to 50% of breast cancers demonstrate clustered microcalcifications. We investigate the potential of using second-order histogram textural features for their correlation with malignancy. A combination of image structure features extracted from the second histogram was used with binary cluster features extracted from segmented calcifications. Several architectures of neural networks were used for analyzing the features. The neural network yielded good results for the classification of hard-to-diagnose cases of mammographic microcalcification into benign malignant categories using the selected set of features. >