Application of artificial neural networks for the classification of liver lesions by image texture parameters
TL;DR: In this article, a multilayered back-propagation neural network was used for liver lesion classification using B-scan ultrasound images for normal, hemangioma and malignant livers.
Abstract: Ultrasound imaging is a powerful tool for characterizing the state of soft tissues; however, in some cases, where only subtle differences in images are seen as in certain liver lesions such as hemangioma and malignancy, existing B-scan methods are inadequate. More detailed analyses of image texture parameters along with artificial neural networks can be utilized to enhance differentiation. From B-scan ultrasound images, 11 texture parameters comprising of first, second and run length statistics have been obtained for normal, hemangioma and malignant livers. Tissue characterization was then performed using a multilayered backpropagation neural network. The results for 113 cases have been compared with a classification based on discriminant analysis. For linear discriminant analysis, classification accuracy is 79.6% and with neural networks the accuracy is 100%. The present results show that neural networks classify better than discriminant analysis, demonstrating a much potential for clinical application.
Citations
More filters
Proceedings Article•
20 Feb 2011
TL;DR: The role of the textural parameters in characterizing different types of inflammatory bowel diseases and the colorectal tumors is studied and the time intensity curves (TIC) are post processed through mathematical modeling, in order to emphasize the behavior of the contrast agent for the considered affections.
Abstract: The inflammatory bowel diseases (IBD) are severe, chronic and recurring disorders, requiring continuous patient monitoring. We aim to develop computerized methods for the noninvasive assessment of the bowel inflammation based on ultrasound images. In this work, we study the role of the textural parameters in characterizing different types of inflammatory bowel diseases and the colorectal tumors. Also, the time intensity curves (TIC) are post processed through mathematical modeling, in order to emphasize the behavior of the contrast agent for the considered affections. The relevant textural and TIC parameters are determined through specific methods, and then they are assessed individually and in combination for performing automatic diagnosis. B-mode and contrast-enhanced ultrasound images (CEUS) of biopsied patients are used. The information obtained from the endoscopic images is taken into consideration, as well. The patients were suffering from the following diseases: Crohn's disease, ulcerative recto-colitis, colon cancer.
1 citations
10 Oct 2008
TL;DR: The textural imagistic model of ADKP is extended by adding new, more expressive textural features and by improving the feature selection methods through combining them in an efficient manner.
Abstract: The prostatic adenocarcinoma (ADKP) is the most frequent neoplasy and also the major cause of death for men in United States. Detecting this tumor by human eye from biomedical images is difficult and invasive methods like the prostate needle biopsy are dangerous for the patient. The aim of our research is to develop reliable, non-invasive, computerized methods in order to provide an accurate characterization of ADKP through textural features extracted from ultrasound images, for the final purpose of automatic diagnosis. Thus, in our previous works, we defined the textural imagistic model of ADKP, consisting in the non-redundant set of the best textural features appropriate for ADKP characterization and in the statistical parameters associated to each relevant feature. In this work, we extend the textural imagistic model of ADKP by adding new, more expressive textural features and by improving the feature selection methods through combining them in an efficient manner.
1 citations
13 Oct 2011
TL;DR: This work analyzes the role that the superior order Gray Level Cooccurrence Matrices (GLCM) and the associated parameters have in the improvement of HCC characterization and automatic diagnosis, and determines the best spatial relation between the pixels that leads to the highest performances.
Abstract: The hepatocellular carcinoma (HCC) is the most frequent malignant liver tumor. The golden standard for HCC diagnosis is the needle biopsy, but this is an invasive, dangerous method. We aim to develop computerized, non-invasive techniques for the automatic diagnosis of HCC, based on information obtained from ultrasound images. The texture is an important property of the internal organs tissue, able to provide subtle information about the pathology. We previously defined the textural model of HCC, consisting in the exhaustive set of the relevant textural features, appropriate for HCC characterization and in the specific values of these features. In this work, we analyze the role that the superior order Gray Level Cooccurrence Matrices (GLCM) and the associated parameters have in the improvement of HCC characterization and automatic diagnosis. We also determine the best spatial relation between the pixels that leads to the highest performances, for the third and fifth order GLCM.
1 citations
01 Jan 2009
TL;DR: The results indicate that the texture features computed from the gray level cooccurrence matrix do not constitute a good candidate to be used in classification and/or diagnosis of liver tissue as normal or cystic, due to homogeneity of these two tissue types within themselves.
Abstract: Images which are obtained in clinical radiology are generally evaluated visually. Some information which is available in the images, but not possible to be seen visually can be useful for diagnosis of some diseases. Cyst hydatid which is a parasitic liver disease is still an important health problem in countries where animal breeding is widespread. In this study, we aimed at producing some objective measures using image analysis, which will be of assistance in the diagnosis of cyst hydatid. The prevalence rate of this condition is relatively high in Turkey. In order to differentiate between regions of liver with cyst hydatid and healthy parenchymal tissues, we have used second order texture features computed from gray level cooccurrence matrix of liver CT images. We have then used these features from the two groups in designing a classifier using probabilistic neural network. Our results indicate that the texture features computed from the gray level cooccurrence matrix do not constitute a good candidate to be used in classification and/or diagnosis of liver tissue as normal or cystic. This must be due to homogeneity of these two tissue types within themselves. 1. Giris |, ( , ), (, ) , ( , ) }
TL;DR: A computer aided classification system has been proposed for classification of mammogram images into normal, benign and cancer classes and indicates that GLCM mean and range features computed at d=1 yield the maximum overall classification accuracy.
Abstract: In the present work, a computer aided classification system has been proposed for classification of mammogram images into normal, benign and cancer classes. The work has been carried out on thirty Digital Database for Screening mammography (DDSM) cases consisting of 10 normal, 10 benign and 10 cancer images. The regions of interest (ROI) have been extracted from the right Medio Lateral Oblique (RMLO) part of the mammogram. We extracted 256×256 pixel size ROI from each case. Texture descriptors based on gray level co-occurrence method by varying the value of inter pixel distance 'd' from 1 to 8 have been used. The SVM classifier has been used for the classification task. The result of the study indicates that GLCM mean and range features computed at d=1 yield the maximum overall classification accuracy of 75% and 65 % respectively.
References
More filters
01 Nov 1973
TL;DR: These results indicate that the easily computable textural features based on gray-tone spatial dependancies probably have a general applicability for a wide variety of image-classification applications.
Abstract: Texture is one of the important characteristics used in identifying objects or regions of interest in an image, whether the image be a photomicrograph, an aerial photograph, or a satellite image. This paper describes some easily computable textural features based on gray-tone spatial dependancies, and illustrates their application in category-identification tasks of three different kinds of image data: photomicrographs of five kinds of sandstones, 1:20 000 panchromatic aerial photographs of eight land-use categories, and Earth Resources Technology Satellite (ERTS) multispecial imagery containing seven land-use categories. We use two kinds of decision rules: one for which the decision regions are convex polyhedra (a piecewise linear decision rule), and one for which the decision regions are rectangular parallelpipeds (a min-max decision rule). In each experiment the data set was divided into two parts, a training set and a test set. Test set identification accuracy is 89 percent for the photomicrographs, 82 percent for the aerial photographic imagery, and 83 percent for the satellite imagery. These results indicate that the easily computable textural features probably have a general applicability for a wide variety of image-classification applications.
20,442 citations
TL;DR: This paper provides an introduction to the field of artificial neural nets by reviewing six important neural net models that can be used for pattern classification and exploring how some existing classification and clustering algorithms can be performed using simple neuron-like components.
Abstract: Artificial neural net models have been studied for many years in the hope of achieving human-like performance in the fields of speech and image recognition. These models are composed of many nonlinear computational elements operating in parallel and arranged in patterns reminiscent of biological neural nets. Computational elements or nodes are connected via weights that are typically adapted during use to improve performance. There has been a recent resurgence in the field of artificial neural nets caused by new net topologies and algorithms, analog VLSI implementation techniques, and the belief that massive parallelism is essential for high performance speech and image recognition. This paper provides an introduction to the field of artificial neural nets by reviewing six important neural net models that can be used for pattern classification. These nets are highly parallel building blocks that illustrate neural net components and design principles and can be used to construct more complex systems. In addition to describing these nets, a major emphasis is placed on exploring how some existing classification and clustering algorithms can be performed using simple neuron-like components. Single-layer nets can implement algorithms required by Gaussian maximum-likelihood classifiers and optimum minimum-error classifiers for binary patterns corrupted by noise. More generally, the decision regions required by any classification algorithm can be generated in a straightforward manner by three-layer feed-forward nets.
7,798 citations
Book•
01 Jan 1974
TL;DR: The objective is to establish an experimental procedure and show direct AFM progression from EMT to EMT using a simple, straightforward, and reproducible procedure.
Abstract: Pathologic basis of disease , Pathologic basis of disease , کتابخانه دیجیتالی دانشگاه علوم پزشکی و خدمات درمانی شهید بهشتی
5,162 citations
01 Apr 1976
TL;DR: In this paper, three standard approaches to automatic texture classification make use of features based on the Fourier power spectrum, on second-order gray level statistics, and on first-order statistics of gray level differences, respectively.
Abstract: Three standard approaches to automatic texture classification make use of features based on the Fourier power spectrum, on second-order gray level statistics, and on first-order statistics of gray level differences, respectively. Feature sets of these types, all designed analogously, were used to classify two sets of terrain samples. It was found that the Fourier features generally performed more poorly, while the other feature sets all performned comparably.
1,379 citations
TL;DR: The gray value distribution of the runs is proposed to be used to define two new features, viz., low gray level run emphasis ( LGRE) and high gray levelrun emphasis ( HGRE).
443 citations