Neural network based approach for anomaly detection in the lungs region by electrical impedance tomography.
TL;DR: The study shows that there is interaction between the size (radius) and conductivity of anomalies and for some combination of these two factors the classification error of neural networks will be very small.
Abstract: In this paper, we have shown a simple procedure to detect anomalies in the lungs region by electrical impedance tomography. The main aim of the present study is to investigate the possibility of anomaly detection by using neural networks. Radial basis function neural networks are used as classifiers to classify the anomaly as belonging to the anterior or posterior region of the left lung or the right lung. The neural networks are trained and tested with the simulated data obtained by solving the mathematical model equation governing current flow through the simulated thoracic region. The equation solution and model simulation are done with FEMLAB. The effect of adding a higher number of neurons to the hidden layer can be clearly seen by the reduction in classification error. The study shows that there is interaction between the size (radius) and conductivity of anomalies and for some combination of these two factors the classification error of neural networks will be very small.
Citations
More filters
TL;DR: This paper proposes a novel reconstruction process for electrical impedance tomography that offers higher quality images and a higher robustness to noise, and significantly reduces the error associated with image reconstruction.
Abstract: Electrical impedance tomography is a modern biomedical imaging method. Its goal is to image the electrical properties of human tissues. This approach is safe for the patient’s health, is non-invasive and has no known hazards. However, the approach suffers from low accuracy. Linear inverse solvers are commonly used in medical applications, as they are strongly robust to noise. However, linear methods can give only an approximation of the solution that corresponds to a linear perturbation from an initial estimate. This paper proposes a novel reconstruction process. After applying a linear solver, the conductivity distribution is post-processed with a nonlinear algorithm, with the aim of reproducing the abrupt change in conductivity at the boundaries between tissues or organs. The results are used to compare the proposed method with three other widely used methods. The proposed method offers higher quality images and a higher robustness to noise, and significantly reduces the error associated with image reconstruction.
48 citations
10 Nov 2007
TL;DR: The approach extends the traditional window-based strategy by using signal-processing techniques to filter out recurring, background fluctuations in resource behavior and has developed a diagnosis technique that uses standard monitoring data to determine which related changes in behavior may cause anomalies.
Abstract: Identifying and diagnosing anomalies in application behavior is critical to delivering reliable application-level performance. In this paper we introduce a strategy to detect anomalies and diagnose the possible reasons behind them. Our approach extends the traditional window-based strategy by using signal-processing techniques to filter out recurring, background fluctuations in resource behavior. In addition, we have developed a diagnosis technique that uses standard monitoring data to determine which related changes in behavior may cause anomalies. We evaluate our anomaly detection and diagnosis technique by applying it in three contexts when we insert anomalies into the system at random intervals. The experimental results show that our strategy detects up to 96% of anomalies while reducing the false positive rate by up to 90% compared to the traditional window average strategy. In addition, our strategy can diagnose the reason for the anomaly approximately 75% of the time.
38 citations
TL;DR: This paper conducts experiments with a self-organizing map neural network that adapts to the structure of a tactile sheet and spatial resolution of the input tactile device; this adaptation is faster and more robust against noise than image reconstruction techniques based on electrical impedance tomography.
Abstract: The sense of touch is considered as an essential feature for robots in order to improve the quality of their physical and social interactions. For instance, tactile devices have to be fast enough to interact in real-time, robust against noise to process rough sensory information as well as adaptive to represent the structure and topography of a tactile sensor itself - i.e., the shape of the sensor surface and its dynamic resolution. In this paper, we conduct experiments with a self-organizing map (SOM) neural network that adapts to the structure of a tactile sheet and spatial resolution of the input tactile device; this adaptation is faster and more robust against noise than image reconstruction techniques based on Electrical Impedance Tomography (EIT). Other advantages of this bio-inspired reconstruction algorithm are its simple mathematical formulation and the ability to self-calibrate its topographical organization without any a priori information about the input dynamics. Our results show that the spatial patterns of simple and multiple contact points can be acquired and localized with enough speed and precision for pattern recognition tasks during physical contact.
23 citations
Cites methods from "Neural network based approach for a..."
...Peng and Mo (2003) [65] EIT || 32 || neighboring method 10−4 mean squared error (MSE) of networks Multilevel BP neural network (MBPNN) (three levels) Minhas and Reddy (2005) [40] EIT || 16 8% limit error Four different RBFNNs, corresponding to four different classifiers, was trained by applying OLSA Adler and Guardo (1994) [43] EIT || 16 0....
[...]
TL;DR: A study of the resulting image distortion, and a method for correcting this lag in situations where the frame rate is insufficient to prevent significant image degradation is proposed.
Abstract: There has been a surge of interest in using electrical impedance tomography (EIT) for monitoring regional lung ventilation, however, EIT is an ill-conditioned problem, and errors/noise in the boundary voltages can have an undesirable effect on the quality of the final image. Most EIT systems in clinical usage use serial data collection hence data used to create a single image will have been collected at different times. This paper presents a study of the resulting image distortion, and proposes a method for correcting this lag in situations where the frame rate is insufficient to prevent significant image degradation. Significant correlation between the standard deviation of the time dependent reciprocity error and time delay dLe between the reciprocal electrode combinations was found for both adult and neonate data. This was reduced when the data was corrected for dLe. Original and corrected data was reconstructed with the GREIT algorithm and visible differences were found for the neonate data. Ideally EIT systems should be run at a frame rate of at least 50 times the frequency of the dominant and interesting physiological signals. Where this is not practical, the intra-frame system timings should be determined and lag corrected for.
15 citations
Cites methods from "Neural network based approach for a..."
...Modelling A FEM mesh of the thorax, incorporating lungs, spine/sternum and heart compartments with realistic conductivities (Minhas & Reddy 2005) was used to model the effect of lag on image reconstruction....
[...]
References
More filters
TL;DR: The authors propose an alternative learning procedure based on the orthogonal least-squares method, which provides a simple and efficient means for fitting radial basis function networks.
Abstract: The radial basis function network offers a viable alternative to the two-layer neural network in many applications of signal processing. A common learning algorithm for radial basis function networks is based on first choosing randomly some data points as radial basis function centers and then using singular-value decomposition to solve for the weights of the network. Such a procedure has several drawbacks, and, in particular, an arbitrary selection of centers is clearly unsatisfactory. The authors propose an alternative learning procedure based on the orthogonal least-squares method. The procedure chooses radial basis function centers one by one in a rational way until an adequate network has been constructed. In the algorithm, each selected center maximizes the increment to the explained variance or energy of the desired output and does not suffer numerical ill-conditioning problems. The orthogonal least-squares learning strategy provides a simple and efficient means for fitting radial basis function networks. This is illustrated using examples taken from two different signal processing applications. >
3,414 citations
TL;DR: A survey of the work in electrical impedance tomography can be found in this article, where the authors survey some of the most important works in the field. Butt.t.
Abstract: t. This paper surveys some of the work our group has done in electrical impedance tomography.
1,726 citations
264 citations
TL;DR: A description is given of an instrument designed to acquire data for the construction of images of internal body structures based on measurements of electrical impedance made from a set of electrodes applied around the periphery of the body.
Abstract: A description is given of an instrument designed to acquire data for the construction of images of internal body structures based on measurements of electrical impedance made from a set of electrodes applied around the periphery of the body. The instrument applies currents at 15 kHz in any desired pattern to 32 electrodes and measures the resulting voltage at each electrode. The construction of a test phantom is also described and the results of initial studies showing the distinguishability of targets of differing sizes and conductivities placed in the phantom are reported. The system is able to distinguish the presence of 9-mm-diameter insulators or conductors placed in the center of a 30-cm-diameter circular tank of salt water. This system is capable of implementing an adaptive process of produce the best currents to distinguish the unknown conductivity from a homogeneous conductivity. >
155 citations
TL;DR: This article focuses on the type of EIT called adaptive current tomography (ACT) in which currents are applied simultaneously to all the electrodes, where a number of current patterns are applied, where each pattern defines the current for each electrode, and the subsequent electrode voltages are measured to generate the data required for image reconstruction.
Abstract: Electrical impedance tomography (EIT) is an imaging modality that estimates the electrical properties at the interior of an object from measurements made on its surface. Typically, currents are injected into the object through electrodes placed on its surface, and the resulting electrode voltages are measured. An appropriate set of current patterns, with each pattern specifying the value of the current for each electrode, is applied to the object, and a reconstruction algorithm uses knowledge of the applied current patterns and the measured electrode voltages to solve the inverse problem, computing the electrical conductivity and permittivity distributions in the object. This article focuses on the type of EIT called adaptive current tomography (ACT) in which currents are applied simultaneously to all the electrodes. A number of current patterns are applied, where each pattern defines the current for each electrode, and the subsequent electrode voltages are measured to generate the data required for image reconstruction. A ring of electrodes may be placed in a single plane around the object, to define a two-dimensional problem, or in several layers of such rings, to define a three-dimensional problem. The reconstruction problem is described and two algorithms are discussed, a one-step, two-dimensional (2-D) Newton-Raphson algorithm and a one-step, full three-dimensional (3-D) reconstructor. Results from experimental data are presented which illustrate the performance of the algorithms.
141 citations