Jan Jantzen

Pap smear diagnosis using a hybrid intelligent scheme focusing on genetic algorithm based feature selection and nearest neighbor classification

The term pap-smear refers to samples of human cells stained by the so-called Papanicolaou method. The purpose of the Papanicolaou method is to diagnose pre-cancerous cell changes before they progress to invasive carcinoma. In this paper a metaheuristic algorithm is proposed in order to classify the cells. Two databases are used, constructed in different times by expert MDs, consisting of 917 and 500 images of pap smear cells, respectively. Each cell is described by 20 numerical features, and the cells fall into 7 classes but a minimal requirement is to separate normal from abnormal cells, which is a 2 class problem. For finding the best possible performing feature subset selection problem, an effective genetic algorithm scheme is proposed. This algorithmic scheme is combined with a number of nearest neighbor based classifiers. Results show that classification accuracy generally outperforms other previously applied intelligent approaches.

An aphasia database on the internet: a model for computer-assisted analysis in aphasiology.

A web-based software model was developed as an example for data mining in aphasiology. It is used for educating medical and engineering students. It is based upon a database of 254 aphasic patients which contains the diagnosis of the aphasia type, profiles of an aphasia test battery (Aachen Aphasia Test), and some further clinical information. In addition, the cerebral lesion profiles of 147 of these cases were standardized by transferring the coordinates of the lesions to a 3D reference brain based upon the ACPC coordinate system. Two artificial neural networks were used to perform a classification of the aphasia type. First, a coarse classification was achieved by using an assessment of spontaneous speech of the patient which produced correct results in 87% of the test cases. Data analysis tools were used to select four features of the 30 available test features to yield a more accurate diagnosis. This classifier produced correct results in 92% of the test cases. The neural network approach is similar to grouping performed in group studies, while the nearest-neighbor method shows a design more similar to case studies. It finds the neurolinguistic and the lesion data of patients whose AAT profiles are most similar to the users input. This way lesion profiles can be compared to each other interindividually. The Aphasia Diagnoser is available on the Web address http://fuzzy.iau.dtu.dk/aphasia.nsf and thus should facilitate a discussion about the reliability and possibilities of data-mining techniques in aphasiology.

Diagnosis of Aphasia Using Neural And Fuzzy Techniques

The language disability aphasia has several sub-diagnoses such as Amnestic, Broca, Global, and Wernicke. Data concerning 265 patients is available in the form of test scores and diagnoses, made by physicians according to the Aachen Aphasia Test. A neural network model has been built, which is available for consultation on the World Wide Web. The neural network model is in this paper compared with a fuzzy model. Rather than concluding which method provides the best approximation, the paper acts as an example solution useful for other benchmark studies.

The application of fuzzy-based methods to central nerve fiber imaging

This paper discusses the potential of fuzzy logic methods within medical imaging. Technical advances have produced imaging techniques that can visualize structures and their functions in the living human body. The interpretation of these images plays a prominent role in diagnostic and therapeutic decisions, so physicians must deal with a variety of image processing methods and their applications. This paper describes three different sources of medical imagery that allow the visualization of nerve fibers in the human brain: (1) an algorithm for automatic segmentation of some parts of the thalamus in magnetic resonance images based on the differences in myelin content in various thalamic subnuclei; (2) polarized light for classifying the 3D orientation of the nerve fibers at each point; and (3) confocal laser scanning microscopy (CLSM) for calculating semiquantitative variables for myelin content. Fuzzy logic methods were applied to analyze these pictures from low- to high-level image processing. The solutions presented here are motivated by problems of routine neuroanatomic research demonstrating fuzzy-based methods to be valuable tools in medical image processing.

Inference planning using digraphs and Boolean arrays

The study applies digraph techniques to a paper on Boolean array structures in rule bases. The rule base is mapped into a digraph, represented by Boolean matrices. Techniques for computing reachability, and for finding the firing order of rules using level codes are given in the Nested Interactive Array Language (Nial). The digraph approach leads to a number of tests for consistency.

Representing knowledge with functions and Boolean arrays

Over the past eighteen years a variety of advanced decision support systems have been built with knowledge-based expert system (KBES) components. For the past eight years, a knowledge representation and manipulation (KRM) scheme called FABA (Functions And Boolean Arrays) has been used. It has two basic principles. First, knowledge is viewed as a functional mapping between input and output variables, where the functions are expressed as fact tables or bases and procedure modules. Second, the function network can be represented with Boolean arrays. The basics of FABA, its implementation in APLS, and a simple example of FABA’s application in a manufacturing dispatch application for IBM’s semiconductor facility in Burlington, Vermont, are described in this paper.

Pap-Smear Classification Using Efficient Second Order Neural Network Training Algorithms

In this paper we make use of two highly efficient second order neural network training algorithms, namely the LMAM (Levenberg-Marquardt with Adaptive Momentum) and OLMAM (OptimizedLevenberg-Marquardt with Adaptive Momentum), for the construction of an efficient pap-smear test classifier. The algorithms are methodologically similar, and are based on iterations of the form employed in the Levenberg-Marquardt (LM) method for non-linear least squares problems with the inclusion of an additional adaptive momentum term arising from the formulation of the training task as a constrained optimization problem. The classification results obtained from the application of the algorithms on a standard benchmark pap-smear data set reveal the power of the two methods to obtain excellent solutions in difficult classification problems whereas other standard computational intelligence techniques achieve inferior performances.

Intelligent and nature inspired optimization methods in medicine: the Pap smear cell classification problem

: The classification problem consists of using some known objects, usually described by a large vector of features, to induce a model that classifies others into known classes. Feature selection is widely used as the first stage of the classification task to reduce the dimension of the problem, decrease noise and improve speed by the elimination of irrelevant or redundant features. The present paper deals with the optimization of nearest neighbour classifiers via intelligent and nature inspired algorithms for a very significant medical problem, the Pap smear cell classification problem. The algorithms used include tabu search, genetic algorithms, particle swarm optimization and ant colony optimization. The proposed complete algorithmic scheme is tested on two sets of data. The first consists of 917 images of Pap smear cells and the second set consists of 500 images, classified carefully by expert cyto-technicians and doctors. Each cell is described by 20 numerical features, and the cells fall into seven classes representing a variety of normal and abnormal cases. Nevertheless, from the medical diagnosis viewpoint, a minimum requirement corresponds to the general two-class problem of correct separation between normal and abnormal cells.

A Fuzzy Control Course on the Internet

A course in fuzzy control has been offered on the Internet (http://www.iau.dtu.dk/~jj/learn) since late 1995. Now that about 100 students have taken the course it is possible to report some results and experiences. The course concerns fuzzy logic for automatic control. The objectives are to teach the basics of fuzzy logic, to show how to use fuzzy logic, and to teach how to design a fuzzy controller. A ball balancer, an inverted pendulum problem, acts as a case study implemented in a software simulator in Matlab. The paper aims at teachers who might be interested in running a similar course, and it presents aspects of the development, the delivery, and the use of the course. The course differs from other distance learning courses in the close teacher-student interaction based on e-mail. It has been set up and run by one person.

Fuzzy Control Tutorial

The tutorial concerns automatic control of an inverted pendulum, especially rule based control by means of fuzzy logic. A ball balancer, implemented in a software simulator in Matlab, is used as a practical case study. The objectives of the tutorial are to teach the basics of fuzzy control, and to show how to apply fuzzy logic in automatic control. The tutorial is distance learning, where students interact one-to-one with the teacher using e-mail.