Jacek J. Dusza

Stereovision for dynamic analysis of human body movement

The aim of this paper is to describe a new Stereovision Laboratory. It was designed and constructed by the authors of the paper in the University of Valencia. It consists of four cameras with its special equipment, a moving track, hardware and software for calculation and visualization. The tests carried out on the measurement stand lead to obtain the set of data describing the movement of human body. The stored data serves to analyze static and dynamic behavior of some parts of human body like arms, legs or vertebral column. Special software was design in order to do visual inspection of the movement and to select a proper data.

Simplified model of mean double step (MDS) in human body movement

In this paper we present a simplified and useful model of the human body movement based on the full gait cycle description, called the Mean Double Step (MDS). It enables the parameterization and simplification of the human movement. Furthermore it allows a description of the gait cycle by providing standardized estimators to transform the gait cycle into a periodical movement process. Moreover the method of simplifying the MDS model and its compression are demonstrated. The simplification is achieved by reducing the number of bars of the spectrum and I or by reducing the number of samples describing the MDS both in terms of reducing their computational burden and their resources for the data storage. Our MDS model, which is applicable to the gait cycle method for examining patients, is non-invasive and provides the additional advantage of featuring a functional characterization of the relative or absolute movement of any part of the body.

SVM detection of premature ectopic excitations based on modified PCA

The paper presents a modified version of principal component analysis of 3-channel Holter recordings that enables to construct one SVM linear classifier for the selected group of patients with arrhythmias. Our classifier has perfect generalization properties. We studied the discrimination of premature ventricular excitation from normal ones. The high score of correct classification (95%) is due to the orientation of the system of coordinates along the largest eigenvector of the normal heart action of every patient under study.

PCA Representation of ECG Signal as a Useful Tool for Detection of Premature Ventricular Beats in 3-Channel Holter Recording by Neural Network and Support Vector Machine Classifier

In the paper classification method of compressed ECG signal was presented. Classification of single heartbeats was performed by neural networks and support vector machine. Parameterization of ECG signal was realized by principal component analysis (PCA). For every heartbeat only two descriptors have been used. The results of real Holter signal were presented in tables and as plots in planespherical coordinates. The efficiency of classification is near to 99%.

Application of binary pulse signals to coincidence methods

Abstract A simple shift-register random-pulse generator is described. The output signal can be applied to verification methods in liquid-scintillation coincidence systems for the direct determination of absolute radioactivity. Some properties of the signals generated are discussed. The method of control of the Double to Single Pulse Coincidence Ratio (DSCR) and the Triple to Double Coincidence Ratio (TDCR) is also presented.

Averaged EMG signal models obtained in cyclic processes

The main aim of this paper is to verify an objective method of averaging electromyographic signals for cyclic processes, which are presented on Symposium Wilga 2016 [1]. The process of creating an objective EMG signal model will be considered. The filtration, the method of smoothing the signal, the sequence of actions and the way they were carried out were taken into account. It is also attempts to measure and analyze the electromyographic signals obtained from the muscles of the lower limbs during the cyclic activity of walking with different loads. This research is based on the data collected in the Laboratory of Movement Analysis at the Medical University of Warsaw. The object of research was a healthy man in the age of 35. Based on the collected data, 130 objective EMG models were created. They correspond to ten muscles of the lower limbs. There were 13 different models calculated for each muscle, corresponding to its activity during the unloaded gait cycle and with left-, right- and double-sided loading for 2,4,6 and 8 kg. The signal models created in this way became the basis for the analysis, as a function of the load, of certain parameters of EMG signals generated during the walk. For each of the EMG signal models, the following parameters were calculated: a moment of muscle activation, duration of muscle activation, physiological costs incurred during muscle activation. Collected data could compere muscle activities and calculate the phases of their interactions.

Objective models of EMG signals for cyclic processes such as a human gait

EMG signals are small potentials appearing at the surface of human skin during muscle work. They arise due to changes in the physiological state of cell membranes in the muscle fibers. They are characterized by a relatively low frequency range (500 Hz) and a low amplitude signal (of the order of μV), making it difficult to record. Raw EMG signal is inherently random shape. However we can distinguish certain features related to the activation of the muscles of a deterministic or quasi-deterministic associated with the movement and its parametric description. Objective models of EMG signals were created on the base of actual data obtained from the VICON system installed at the University of Physical Education in Warsaw. The object of research (healthy woman) moved repeatedly after a fixed track. On her body 35 reflective markers to record the gait kinematics and 8 electrodes to record EMG signals were placed. We obtained research data included more than 1,000 EMG signals synchronized with the phases of gait. Test result of the work is an algorithm for obtaining the average EMG signal received from the multiple registration gait cycles carried out in the same reproducible conditions. The method described in the article is essentially a pre-finding measurement data from the two quasi-synchronous signals at different sampling frequencies for further processing. This signal is characterized by a significant reduction of high frequency noise and emphasis on the specific characteristics of individual records found in muscle activity.

Asymmetry functions of joint angles in the human gait

The idea of this project is to study the asymmetry during human walking on some important angles. For this purpose, asymmetry functions have been developed. Until this moment, the asymmetry on these angles has been studied like something global with the asymmetry coefficients (for example, just taking the best value of asymmetry for every angle). With the asymmetry functions it is possible to study this asymmetry along the whole. Knowing this information, it could be possible to determine different anomalies depending on the angle and on the moment of the asymmetry. The ultimate goal is to obtain a computer program to perform the asymmetry functions at different angles from data obtained from patients.

Symmetrical gait descriptions

The article presents a proposal for the description of human gait as a periodic and symmetric process. Firstly, the data for researches was obtained in the Laboratory of Group SATI in the School of Engineering of University of Valencia. Then, the periodical model – Mean Double Step (MDS) was made. Finally, on the basis of MDS, the symmetrical models – Left Mean Double Step and Right Mean Double Step (LMDS and RMDS) could be created. The method of various functional extensions was used. Symmetrical gait models can be used to calculate the coefficients of asymmetry at any time or phase of the gait. In this way it is possible to create asymmetry, function which better describes human gait dysfunction. The paper also describes an algorithm for calculating symmetric models, and shows exemplary results based on the experimental data.

Simplification of the kinematic model of human movement

The paper presents a methods of simplification of the human gait model. The experimental data were obtained in the laboratory of the group SATI in the Electronics Engineering Department of the University of Valencia. As a result of the Mean Double Step (MDS) procedure, the human motion were described by a matrix containing the Cartesian coordinates of 26 markers placed on the human body recorded in the 100 time points. With these data it has been possible to develop an software application which performs a wide diversity of tasks like array simplification, mask calculation for the simplification, error calculation as well as tools for signals comparison and movement animation of the markers. Simplifications were made by the spectral analysis of signals and calculating the standard deviation of the differences between the signal and its approximation. Using this method the signals of displacement could be written as the time series limited to a small number of harmonic signals. This approach allows us for a high degree of data compression. The model presented in this work can be applied into the context of medical diagnostics or rehabilitation because for a given approximation error and a large number of harmonics may demonstrate some abnormalities (of orthopaedic symptoms) in the gait cycle analysis.