Simultaneous and sequential same-arm measurements in the validation studies of automated blood pressure measuring devices.

The oscillometric ambulatory blood pressure recorder Daypress 500 was validated according to the British Hypertension Society protocol. Both sequential and simultaneous measurements were used. Multiple regression analysis demonstrated a significant influence of subject pulse pressure and arm circumference on device-observer systolic pressure differences. Differences between observer consecutive readings were inversely related to heart rate. Device and observer blood pressure readings were closer at simultaneous than at sequential measurements. However, both kinds of measurement led to the same final evaluation (A for diastolic and B for systolic blood pressure), provided that the appropriate grading criteria were applied for each method.

Integrating content-based retrieval in a medical image reference database.

Image reference databases (IRDBs) are a recurrent research topic in medical imaging. Most IRDBs are designed to help experienced physicians in diagnostic tasks and require that users have prior extensive knowledge of the field for their use to be fruitful. Therefore, the educational potential of such image collections cannot be exploited thoroughly. In this paper we propose an image-indexing method to extend the functionalities of an existing medical IRDB and allow for its use in educational applications, as well as in computer-assisted diagnosis. Our method, based on the Kahrunen-Leève transform, has been used to develop a content-based search engine for tomographic image databases on which we are presently experimenting and which we aim to integrate into a working radiological IRDB installed at the University of Florence. Results achieved in our preliminary tests are also reported.

Genetic design of optimum linear and nonlinear QRS detectors.

This paper describes an approach to the design of optimum QRS detectors. We report on detectors including a linear or nonlinear polynomial filter, which enhances and rectifies the QRS complex, and a simple, adaptive maxima detector. The parameters of the filter and the detector, and the samples to be processed are selected by a genetic algorithm which minimizes the detection errors made on a set of reference ECG signals. Three different architectures and the experimental results achieved on the MIT-BIH Arrhythmia Database are described.

Neural network segmentation of magnetic resonance spin echo images of the brain.

This paper describes a neural network system to segment magnetic resonance (MR) spin echo images of the brain. Our approach relies on the analysis of MR signal decay and on anatomical knowledge; the system processes two early echoes of a standard multislice sequence. Three main subsystems can be distinguished. The first implements a model of MR signal decay; it synthesizes a four-echo multiecho sequence, in order to add images characterized by long echo-times to the input sequence. The second subsystem exploits a priori anatomical knowledge by producing an image, in which pixels belonging to brain parenchyma are highlighted. Such anatomical information allows the following submodule to distinguish biologically different tissues with similar water content, and hence similar appearance, which might produce misclassifications. The grey levels of the reconstructed sequence and the output of the second module are processed by the third subsystem, which performs the segmentation of the sequence. Each pixel is assigned to one of five different tissue classes that can be revealed with brain MR spin echo imaging. With a suitable encoding, a five-level segmented image can then be produced. The system is based on feed-forward networks trained with the back-propagation algorithm; experiments to assess its performance have been carried out on both simulated and clinical images.

Neural network synthesis of spin echo multiecho sequences.

Spin echo multiecho sequences are not frequently used in clinical practice, because they allow the observation of one single slice, imaged at different echo times, for each acquisition. To limit examination time, multislice sequences that include only images derived from one or two echoes are usually acquired. Nevertheless, the strong T2 dependence of multiecho sequences can be used effectively to enhance the contrast between tissues with different T2 and to gather useful diagnostic information. Artificial neural networks can offer new interesting facilities to the radiologist. In fact, the learning capabilities of neural networks allow them to extract the prototypical behavior of a system from a set of examples. After learning, artificial neural networks can emulate the system behavior even in the presence of new inputs, as far as these are not too different from those included in the training set. A conveniently trained neural network can synthesize a multiecho sequence for each slice of a multislice sequence, requiring only two images for each slice to achieve reliable results. When compared with a true multiecho sequence, the images generated by the network preserve the contrast characteristics of the original ones and have a better signal-to-noise (SNR) ratio. In this paper we report the results achieved by using a neural network to reconstruct synthetic spin echo multiecho images of the brain.

Knowledge-based system for the diagnosis and treatment of hypertension.

A knowledge-based system to assist the physician in the diagnosis and treatment of hypertension has been developed as the result of cooperation between the Department of Electronic Engineering of the University of Florence and the Interuniversity Centre of Clinical Chronobiology. The system input consists of the data recorded over a 24 h (or longer) period by monitoring (automatically or through self-measurements) the blood pressure of the subject undergoing the system analysis, and the associated anamnestic data. The process results in a report that states from which kind of hypertensive syndrome, if any, the subject is suffering and which anti-hypertensive therapy appears to be most suitable. The system consists of three modules: the first diagnoses hypertension by applying cluster analysis to a set of parameters derived from the principal components of the time series resulting from the subject's blood pressure monitoring; the other two classify hypertension and offer advice about the most advisable treatment, respectively, by using high-level data representation and processing. The knowledge embedded in the system is internally represented by means of frames and rules. This paper describes the structure of the system, illustrates the techniques that have been used for its development and discusses the results of its application.