Computer assisted characterization of diffused liver disease using image texture analysis techniques on B-scan images (original) (raw)
Related papers
Evaluation of Texture Features in Hepatic Tissue Characterization from Non-enhanced CT Images
2007
Aim of this paper is to evaluate the diagnostic contribution of various types of texture features in discrimination of hepatic tissue in abdominal non-enhanced Computed Tomography (CT) images. Regions of Interest (ROIs) corresponding to the classes: normal liver, cyst, hemangioma, and hepatocellular carcinoma were drawn by an experienced radiologist. For each ROI, five distinct sets of texture features are extracted using First Order Statistics (FOS), Spatial Gray Level Dependence Matrix (SGLDM), Gray Level Difference Method (GLDM), Laws' Texture Energy Measures (TEM), and Fractal Dimension Measurements (FDM)
Non-invasive, image based detection of diseases is one of the most important issues in the nowadays research of biomedical images, because it prevents from some serious problems, that could be generated by the invasive techniques and could be dangerous for the patients. Texture is a fundamental visual property of the tissue providing a lot of information concerning its pathological state. Thus, we developed specific methods for texture analysis and recognition, for automatic and semi-automatic detection of some liver diseases from ultrasound images, in order to assist the medical personal in establishing a diagnostic in non-invasive way. We also performed some studies concerning the relevance of these parameters in the case of various liver diseases.
2008
This paper is focused on current progress of our research in improving diagnosis value of ultrasound imaging in the context of diffuse liver diseases. Image features are computed on ultrasound images and these features are used to train a classifier. The classifier is able to distinguish between various pathology grades. Present study shows that, based on ultrasound images, steatosis can be accurately graded and a qualitative assessment can be made in case of fibrosis. Further improvements can be made if we include more patients and consider non-imagistic features like clinical and biochemical analysis of the patient.
Texture Analysis for Liver Segmentation and Classification: A Survey
2011 Frontiers of Information Technology, 2011
Texture is a combination of repeated patterns with regular/irregular frequency. It can only be visualized but hard to describe in words. Liver structure exhibit similar behavior; it has maximum disparity in intensity texture inside and along boundary which serves as a major problem in its segmentation and classification. Problem gets more complicated when one applies simple segmentation techniques without considering variation in intensity texture. The problem of representing liver texture is solved by encoding it in terms of certain parameters for texture analysis. Numerous textural analysis techniques have been devised for liver classification over the years some of which work equally work well for most of the imaging modalities. Here, we attempt to summarize the efficacy of textural analysis techniques devised for Computed Tomography (CT), Ultrasound and some other imaging modalities like Magnetic Resonance Imaging (MRI), in terms of well-known performance metrics.
users.utcluj.ro
Texture analysis is viewed as a method to enhance the diagnosis power of classical B-mode ultrasound image. Present study aims to evaluate the dependence between the human expert and the performance of such a texture analysis system in predicting the cirrhosis in chronic hepatitis C patients. 125 consecutive chronic hepatitis C patients were included in this study. All the patients had positive HCV-RNA in serum and had undergone percutaneous liver biopsy for disease staging using Metavir score. Ultrasound images were acquired from each patient and 4 experts established regions of interest. Textural analysis software generated 234 features from each region of interest. Relevant textural features were identified and a classification schema was evaluated. Texture analysis can discriminate between F0 and F4 fibrosis stages (AUROC=0.64). The performance of this approach depends highly on the human expert that establishes the regions of interest (p<0.05). The relevant textural features were identified and it was shown that the detection performance didn't depend on the particular feature selection (p=0.8). In classical form met in literature non invasive diagnosis through texture analysis has limited utility in clinical practice because of the user variability introduced by the expert who establishes the regions of interest.
Texture Analysis of Cirrhosis Liver using Support Vector Machine
2014
Diagnostic ultrasound is a useful and noninvasive method in clinical medicine. Although due to its qualitative, subjective and experience-based nature, ultrasound image interpretation can be influenced by image conditions such as scanning frequency and machine settings. In this paper, a method is proposed to extract the cirrhosis and normal liver features using the entropy of texture edge co-occurrence matrix based on ultrasound images, which is not sensitive to changes in emission frequency and gain. Then, support vector machine are employed to test a group of 30 in-vivo liver cirrhosis images from 18 patients, as well as other 30 liver images from 18 normal human volunteers. The results showed that the support vector machine is 94.4% in sensitivity for liver cirrhosis (LC) while neural network provided 92.31 % and the system is considered to be helpful for clinical and educational use.
Journal of Clinical Ultrasound, 1991
By means of statistical pattern recognition procedures, a quantitative description of the ultrasound B-scan images of experimental diffuse liver disease has been carried out. Fatty livers, fatty fibrosiskirrhosis, and cirrhosis without fatty infiltration of the liver were studied in female Wistar rats. Separation accuracies of more than 80% between the tissue classes "normal" vs "fatty infiltration," or "normal" vs "fatty cirrhosis," using only two statistical image parameters were found. A subclassification of the diffuse parenchymal liver disease was not possible. It is shown by multiple linear regression analysis that the image parameter "mean grey level" correlates better with total lipid content than with the amount of connective tissue. Furthermore it is demonstrated that connective tissue leads only to a weak increase in "mean grey level," whereas the addition of connective tissue to a given tissue lipid can lead to a reduction in image brightness.
Texture characterization for hepatic tumor recognition in multiphase CT
Biocybernetics and …, 2006
A new approach to texture characterization from dynamic CT scans of the liver is presented. Images with the same slice position and corresponding to three typical acquisition phases are analyzed simultaneously. Thereby texture evolution during the propagation of contrast product is taken into account. The method is applied to recognizing hepatic primary tumors. Experiments with various sets of texture parameters and two classification methods show that simultaneous analysis of texture parameters derived from three subsequent acquisition moments improves the classification accuracy.
International Journal of Biomedical Engineering and Technology, 2007
Image analysis techniques have played an important role in several medical applications. In this paper, the classification of ultrasonic liver images is studied by using texture features extracted from Laws' method, Autocorrelation method, Edge frequency methods, Gabor Wavelet method and Co-occurrence probability method. Then the best features from different methods are combined to improve the classification. The features from these methods are used to classify four sets of ultrasonic liver images-Normal, Cyst, Benign and Malignant, and how well they suit in classifying the abnormalities is reported. A Neural Network classifier is employed to evaluate the performance of these features based on their recognition ability.
Prediction of liver cirrhosis based on multiresolution texture descriptors from B-mode ultrasound
International Journal of Convergence Computing, 2013
A computer aided diagnostic system to characterise normal and cirrhotic liver by multiresolution texture descriptors is proposed in this paper. The study is carried out in 120 segmented regions of interest extracted from 31 clinically acquired B-mode liver ultrasound images. Mean and standard deviation multiresolution texture descriptors derived by using 2D-discrete wavelet transform, 2D-wavelet packet transform and 2D-Gabor wavelet transform are considered for analysis and exhaustive search with J 3 criterion of class separability is used for feature selection. The performance of subset of five most discriminative texture descriptors obtained from 2D-discrete wavelet transform, 2D-wavelet packet transform and 2D-Gabor wavelet transform is compared by using a support vector machine classifier. It is observed that only five mean multiresolution texture descriptors obtained from 2D-Gabor wavelet transform at selective scale and orientations provide highest classification accuracy of 98.33% and sensitivity of 100% by using a support vector machine classifier. The promising results indicate that the selective frequency and orientation properties of Gabor filters are extremely useful for providing multiscale texture description.