The success of in-vitro fertilization can be predicted by a correct quantitative and qualitative assessment of ovarian follicles. Several ovarian follicle detection and recognition algorithms have been published. Their effectiveness is inferior to human follicle annotations due to various kinds of noise, degradations, and artefacts in ultrasonic images. This paper deals with an approach to recognize antral follicles from 2 mm in diameter in 3D ultrasound data. Its detection phase looks for candidate follicular regions, while the recognition phase assesses the likelihood of a region to correspond to a follicle. Three innovative definitions underpin the detection: Laplacian-of-Gaussian-based directional 3D wavelet transform, adaptive multiscale search based on Gaussian mixtures, and recursive convexity-based region splitting. A likelihood index is also introduced to support follicle recognition. The proposed approach was tested on 30 ultrasound ovarian volumes generated by different sonographic machines in stimulated and non-stimulated examination cycles. The obtained follicle recognition rates exceed those of the best 3D approaches known by about 10 percent, while qualitative assessments yield comparable values.

This paper deals with an optimized heartbeat detection multimethod by using maximum a posteriori probability (MAP). The approach was derived for unobtrusive fiber-optic measurements of cardiac activity. Multiple independent detection methods were selected and characterized by their delays and variability referring to cardiac electrical excitation (R waves). Validation of the approach was performed in two experiments involving 24 participants: 10 interferometric signals were recorded at rest, 14 with variable heart rate after physical exercise. The proposed MAP heartbeat detection was assessed by a cross validation in 250 iterations. Obtained results show the overall efficiency, which was estimated by a product of the sensitivity, precision, and variability of heartbeat detections, yields 97.04 ± 3.36% for the experiment with physical exercise and 97.07 ± 4.49% at rest. The method's accuracy guarantees that the heartbeat detections differ for 22 ± 5 ms and 22 ± 3 ms from the ECG reference in the two types of experiments, respectively.

Various applications of cellular neural networks (CNNs) on complex image processing tasks raise questions about an appropriate selection of template elements that determine the CNN's behavior. In previous research we utilized multiple time variant template elements cellular neural networks for segmentation, which has many advantages compared to conventional search space approaches. In this paper a novel approach which utilizes the formalism of support vector machines (SVMs) that utilizes only single time invariant set of CNN template elements, is introduced. The main advantage of this approach is reduction of the number of CNNs templates and confirms to the conventional applications of CNN.

Segmentation of ovarian ultrasound images using cellular neural networks (CNNs) is studied in this paper. The segmentation method consists of five successive steps where the first four uses CNNs. In the first step, only rough position of follicles is determined. In the second step, the results are improved by expansion of detected follicles. In the third step, previously undetected inexpressive follicles are determined, while the fourth step detects the position of ovary. All results are joined in the fifth step. The templates for CNNs were obtained by applying genetic algorithm. The segmentation method has been tested on 50 ovarian ultrasound images. The recognition rate of follicles was around 60% and misidentification rate was around 30%.

Simulated annealing is no doubt one of the most popular optimisation methods. Like similar methods it is based on natural phenomena. In the paper two examples of how to use this method in medical applications are presented. In the first example the successfulness of simulated annealing is compared to a genetic algorithm on the problem of the optimal coefficients determination for cellular neural networks, when used for segmentation of ovarian ultrasound images. In this field the latter method produced better results. The second example discusses the usage of simulated annealing for optimisation of signal classification.

Medicine is a difficult thing to learn. Experimenting with real patients should not be the only option - simulation deserves special attention. The Virtual Reality Modelling Language (VRML), as a tool to build virtual objects and scenes, has a good record of educational applications in medicine, especially for static and animated visualisations of body parts and organs. However, to create computer simulations resembling situations in real environments, the demanded level of interactivity and dynamics is difficult to achieve. In this paper, we describe some approaches and techniques which we used to push the limits of the current VRML technology further towards a dynamic 3D representation of virtual environments. Our demonstration is based on the implementation of a virtual baby model, whose vital signs can be controlled from an external Java application. The main contribution of this work is a proposal for a modified VRML timer/sensor implementation, which greatly improves the overall control of system performance.

Deals with the problem of decomposition of surface electromyograms (SEMG). According to the physiological facts, a multiple-input multiple-output (MIMO) is used. The measured signals are taken as the channel responses corresponding to the motor-unit action potentials (MUAPs) convolution by the innervation pulse trains. The decomposition is based on the third-order cumulants whose values enter as coefficients of nonlinear system of equations. The system is solved by nonlinear least mean square (LMS) optimisation. Synthetic SEMG signals from a MIMO(2,3) with additive Gaussian noise with SNRs of 10 and 0 dB prove that a successful multichannel decomposition is possible also in very noisy environments.

The paper deals with the problem of processing and interpretation of clinically recorded ultrasound images for the reason of following the growth of dominant ovarian follicles in a day-to-day manner. A part of the XUltra project achievements is presented. We propose three different automatic computer-based follicle identification algorithms. The first one is based on cellular neural networks. The second one is based on region growing segmentation method, while the third one processes entire image sequence with a predictor-corrector recognition scheme. The recognition rate of follicles with these algorithms goes up to 78%, while the misidentification rate is around 15%.

The paper describes a higher-order statistics (HOS) approach feasible for decomposition of compound signals. Our novel method introduces asymptotically exact interpolation-based computation of bicepstra with no aliasing. We analysed it in order to establish the conditions under which surface electromyograms (SEMGs) could be decomposed onto their building components, i.e. motor-unit action potentials. The experiments with synthetic SEMGs showed that the results depend highly on the type of the action potentials (APs) respected. Using our novel bicepstral decomposition, the first-norm error for the dipole-based APs falls under 50% of the mean absolute value of the signal samples only with the signals of lengths of over 110000 samples and interpolation level of 4096. The developed decomposition approach be considered suitable for further HOS-based decomposition of SEMGs.

Analyses the scope of multimodal parametric searching (MPS) as a means for the decomposition of electromyograms (EMGs). The approach is based on a predefined set of signal sources that are described by some parameters. These signal sources are searched for in a superimposed signal by calculation of the Euclidian distance and by looking for its minimum. The signal sources are applied successively, according to the changes in their parameters, and peeled off at the minimum Euclidian distance found. The minima detected are then displayed in parametric space as a parametric representation. We show that such a technique is reliable enough for the decomposition of needle EMGs at low contraction forces. Applied to surface EMGs, it still discerns different phenomena, although not always with a direct relationship to reality. This is because of the method's ambiguity, amplitude ratio dependency and a bias in the distribution of the detected signal components. Finally, we introduce a novel approach based on the wavelet transform and using the elements of the parametric search. Both approaches give comparable decomposition results when applied to the EMGs.

Monitoring follicles is especially important in human reproduction. Today, the monitoring of follicles is done non-automatic, with human interaction. This work can be very demanding and inaccurate, and in most cases means only an additional burden for the experts. In this paper, new algorithm for automated computer-assisted defection of follicles in ultrasound images of ovary is proposed. It has typical object recognition scheme (preprocessing, segmentation and classification). The algorithm is assembled on the following idea: first, the ovary is estimated (coarse) and then follicles are searched. The methods used are known from literature (despeckle filter, Kirsh's operator, optimal thresholding, thinning shape descriptions), the majority of the work was done experimenting with these methods and selecting the appropriate thresholds. The algorithms computational complexity is of order O(n/sup 2/), which means about 6 minutes of processing time per ultrasound image of dimensions of 768/spl times/576 pixels (on HP 715 machines). Algorithm is not perfect, but it can be easily modified and improved. Recognition rate of follicles with these procedure is around 70%.

We developed a cellular automaton that is able to extract the information about the position of follicles in the ultrasonic image. Follicles are small protective sacs containing ova in the cortex of women's ovaries, and their ultrasound monitoring is very important in diagnosing and treatment of infertility. The original ultrasonic images were reduced in size and number of colours and then used as the initial state of our cellular automaton. Following the fundamental natural principle of attacking/defending, the transition rule has been developed. The results are quite encouraging and provide a good starting point for further enhancement and development.

A computer program for simulation of electromyographic (EMG) signal, which represents the electric activity of a muscle during contraction, has been built. Some of the simulation models developed in this field were used in its realization. The simulator consists of three major parts: muscle model generation, muscle fiber and motor unit action potential calculation and EMG signal generation. All the parts are controlled by wide range of parameters. In its present state the simulator is capable of simulating three recording techniques: single fiber needle, concentric needle and surface recording technique. The applied simulation models has been proved to agree with the reality and so does, therefore, the generated EMG signal. The simulator has been built to be used as a verification tool for various EMG signal decomposition algorithms, and as a learning tool for young electromyographers on how different muscle parameter settings and recording techniques influence the recorded EMG signal.