A system and method for generating an image representation are provided. The image is modeled as a set of mixture weights, one for each of a set of reference image models, such as Gaussian mixture models (GMMs). The weights are derived by optimizing an objective function in which each reference image model is associated with its respective weight.

Modeling images as mixtures of image models

https://www.hal.inserm.fr/inserm-00411983/document

A vectorial image soft segmentation method based on neighborhood weighted Gaussian mixture model.

The CT uroscan contains three to four time-spaced acquisitions of the same patient. Registration of these acquisitions forms a vectorial volume, which contains a more complete anatomical information. In order to outline the anatomical structures, multi-dimensional classification is necessary for analyzing this vectorial volume. Because of the partial volume effect (PVE), probability distributions are assigned to the different material types within this vectorial volume instead of a definite material distribution. Gaussian mixture model is often used in probability classification problems to model such distributions, but it relies only on the intensity distributions, which will lead a misclassification on the boundaries and inhomogeneous regions with noises. In order to solve this problem, a neighborhood weighted Gaussian mixture model is proposed in this paper. Expectation Maximization algorithm is used as optimization method. The experiments demonstrate that the proposed method can get a better classification result and less affected by the noise.

https://www.hal.inserm.fr/inserm-00316890/document

A vectorial image classification method based on neighborhood weighted Gaussian mixture model

Medical imaging has significantly influenced the field of image processing in recent years. Several advancements have been made in medical imaging that is now being used, and all of them are geared at making disease diagnosis more accurate. Researchers recently explored how to diagnose, define, and classify bone fractures, and numerous approaches have been proposed. However, a uniform categorization has yet to be developed for all detected fractures. This article focuses on the methods for identifying realignments among the bone structures utilizing model-based approaches, and this methodology is highlighted here. The performance metrics are used to assess the findings, and the constructed model has shown substantial results in binary classification and multi-classification of bone fractures. The model that has been suggested is called the Finite Beta Gaussian Mixture Model (FBGMM), and its performance may be evaluated with the use of a confusion matrix. This project intends to construct an image processing system that can identify bone fractures promptly and reliably by including data from X-rays. FGBMM achieves good accuracy in both binary and multiclassification of fractures. 

A statistical model approach based on the Gaussian Mixture Model for the diagnosis and classification of bone fractures

Liver disease is also considered amongst one of the deadly disease for example, Liver cirrhosis (LC), the end phase of numerous types of ceaseless hepatitis of various etiologies is a diffuse procedure portrayed by fibrosis and the change of ordinary liver design into fundamentally abnormal nodules surrounded by annular fibrosis As of now there is one of the predominant illnesses of 21st century is liver issue every year executing such a significant number of individuals' round the universes. The scope of treatments has been given by the analyst to assess results. Early determination is of extensive measure of hugeness in treating the malady. Despite all the standardization methods in medical diagnosis, a correct diagnosis is still considered to be an art much of this situation is for, that medical diagnosis needs proficiency as well as experience in dealing with uncertainty. In spite of the fact that, in our mechanized age, limits of restorative science have very extended, you can't conquer this vulnerability effectively. Offering a groundbreaking structure to develop the model of existing frameworks makes fuzzy hypothesis change to an important factor towards medicinal analysis improvement. Using various Artificial Intelligence methods for liver disorders diagnosis has recently become a wide-spreading area of research. These intelligent systems help physicians as diagnosis assistants. Presently, different Artificial Neural Network framework, Expert Systems, Fuzzy Neural Network and Classification, This book chapter gives an audit of various Artificial System and master framework strategy in determination and identifications of liver infection issue intensity is the key for results. Inspired by these, the work displayed in this paper is centered on contrasting these two incessant liver recognition systems and investigate the outcomes for future research. In the exhibited work, various patients’ datasets are considered are investigated for the location of Liver disease. Moreover, various performance parameters are evaluated like specificity, sensitivity, and accuracy for the overall assessment of the presented model.

A Comprehensive Review of Medical Expert Systems for Diagnosis of Chronic Kidney Diseases

One of the goal of the Nephron-Sparing surgery properative planning is to delineate as exactly as possible the renal carcinoma and to specify its relations to the renal arterial, venous and collecting system anatomies. The classical preoperative imaging system is the spiral CT urography, which gives sucessive 3D acquisitions of complementary information The integration of this information within the a patient specific anatomical referential can be achieved by intra-patient registration techniques. A local MI maximization registration method is proposed in this paper. The kidneys are extracted from the abdomen volumes and then the registration between the extracted kidneys is implemented by maximizing the MI between them. The experimental results demonstrates that this method is effective.

https://www.hal.inserm.fr/inserm-00187319/document

Hui Tang

Papers

Intra subject 3D/3D Kidney Registration using Local Mutual Information Maximization

A vectorial image classification method based on neighborhood weighted Gaussian mixture model