Browsing by Author "Choi, Charles T. M."

Now showing 1 - 20 of 24

A Mesh-Refinement Method Based on Artificial Neural Networks for Electrical Impedance Tomography
(2016-01-01)
Martin, Sebastien
;
Choi, Charles T. M.
;
交大名義發表
;
National Chiao Tung University
Inverse boundary problem are usually solved with a finite element model, with either limited accuracy or huge computation resources. Using a fine mesh results in a computationally demanding task, but when the region of interest is known, when one can easily locally refine the model, aiming for greater local accuracy. In this paper, a novel approach uses artificial neural network to estimate the location of the region of interest and refine this region before solving the inverse problem. The idea is illustrated by solving the electrical impedance tomography inverse problem. Result shows that the proposed method increases the accuracy without significantly affecting the computation resources necessary to solve the inverse problem.
A New Divide-and-Conquer Method for 3-D Electrical Impedance Tomography
(2018-03-01)
Martin, Sebastien
;
Choi, Charles T. M.
;
分子醫學與生物工程研究所
;
電機工程學系
;
Institute of Molecular Medicine and Bioengineering
Electrical impedance tomography (EIT) is a patient-safe approach for imaging applications and is a promising technology for biomedical imaging. By applying a small electrical current to living tissue and measuring the electrical potential at different points of the tissue's boundary, it is possible to solve the inverse problem and to generate a map of the tissue's conductivities. Although it is a promising method, the reconstruction process remains a complicated task, and real-time nonlinear image reconstruction algorithms have not yet been widely adopted. Recent studies have proposed the use of artificial neural networks (ANN) to solve the EIT inverse problem. ANNs can give a nonlinear conductivity distribution. Although several studies have examined 2-D finite-element models, very little work has been done on 3-D problems. In this paper, a solution based on the divide-and-conquer method and ANNs is used to solve the nonlinear inverse problem without any linearization. The solution presented here reduces the difficulty of training large ANNs, which are commonly required to solve a 3-D EIT problem with artificial intelligence algorithms.
A novel post-processing scheme for two-dimensional electrical impedance tomography based on artificial neural networks
(2017-12-05)
Martin, Sebastien
;
Choi, Charles T. M.
;
分子醫學與生物工程研究所
;
資訊工程學系
;
電機工程學系
Objective Electrical Impedance Tomography (EIT) is a powerful non-invasive technique for imaging applications. The goal is to estimate the electrical properties of living tissues by measuring the potential at the boundary of the domain. Being safe with respect to patient health, non-invasive, and having no known hazards, EIT is an attractive and promising technology. However, it suffers from a particular technical difficulty, which consists of solving a nonlinear inverse problem in real time. Several nonlinear approaches have been proposed as a replacement for the linear solver, but in practice very few are capable of stable, high-quality, and real-time EIT imaging because of their very low robustness to errors and inaccurate modeling, or because they require considerable computational effort. Methods In this paper, a post-processing technique based on an artificial neural network (ANN) is proposed to obtain a nonlinear solution to the inverse problem, starting from a linear solution. While common reconstruction methods based on ANNs estimate the solution directly from the measured data, the method proposed here enhances the solution obtained from a linear solver. Conclusion Applying a linear reconstruction algorithm before applying an ANN reduces the effects of noise and modeling errors. Hence, this approach significantly reduces the error associated with solving 2D inverse problems using machine-learning-based algorithms.
A Post-Processing Method for Three-Dimensional Electrical Impedance Tomography
(2017-08-03)
Martin, Sebastien
;
Choi, Charles T. M.
;
分子醫學與生物工程研究所
;
電機工程學系
;
Institute of Molecular Medicine and Bioengineering
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.
A Vocoder for a Novel Cochlear Implant Stimulating Strategy Based on Virtual Channel Technology
(2009)
Choi, Charles T. M.
;
Hsu, C. H.
;
Tsai, W. Y.
;
Lee, Yi Hsuan
;
資訊工程學系
Cochlear implant provides opportunity for profoundly hearing impairment patient to have a chance to hear sound again. However, the limited number of electrodes is in sufficient to provide enough hearing resolution for hearing impaired people, especially for application in tonal language and music. Virtual channel technology opens up the possibility to increase the hearing resolution under the limited electrodes available and improve and quality for tonal language and music. In this paper, vocoder implementation of a new speech strategy based on virtual channel technology is used to study the improvement. The test result shows improved understanding and quality with virtual channels over the traditional strategies, ex. CIS, especially for mandarin and music.
Channel Interaction in Cochlear Implant Acoustic Models
(2016-01-01)
Choi, Charles T. M.
;
Huang, Shang-Yi
;
Lee, Yi-Hsuan
;
分子醫學與生物工程研究所
;
電機工程學系
Cochlear implants (CI) provide opportunities for people with profound hearing impairment to recover partial hearing. An acoustic CI model based on specific stimulating strategy can be used to study sounds processed in CI, and allows test subjects with normal hearing to evaluate stimulating strategy performance. However, there is still a significant performance difference between hearing test results based on acoustic CI models and from CI users. In this study we propose to use a SPREAD matrix, which incorporates channel interaction, created by the activating function profile based on finite element models of CI to improve the acoustic CI model. According to test results, acoustic CI models with SPREAD matrix based on finite element models significantly improves its match with clinical CI data and can be used to evaluate the performance of CI stimulating strategies and predict the hearing performance of CI users more accurately.
Conditions for Generating Virtual Channels in Cochlear Prosthesis Systems
(2009-03-01)
Choi, Charles T. M.
;
Hsu, Chien-Hua
;
分子醫學與生物工程研究所
;
資訊工程學系
;
Institute of Molecular Medicine and Bioengineering
Simultaneous electrical stimulation of neighboring electrodes in cochlear prosthesis systems generates channel interaction. However, intermediate channels, or virtual channels between the neighboring electrodes can be created through controlled channel interaction. This effect may be exploited for sending new information to the hearing nerves by stimulating in a suitable manner. The actual stimulation sites are therefore not limited to the number of electrodes. Clinical experiments, however, show that virtual channels are not always perceived. In this paper, electrical simulation with finite element analysis on a half turn human cochlea model is adopted to model the virtual channel effect, and the conditions for generating virtual channels are discussed. Five input current ratios (100/0, 70/30, 50/50, 30/70, 0/100) are applied to generate virtual channels. Three electrode arrays parameters are taken into consideration: distance between electrode contact and modiolus, spacing between adjacent electrode contacts and scale of electrode contact size. By observing the activating function contours, the virtual channel patterns and performances can be measured and examined. The results showed that a broad excitation pattern is necessary to produce the kind of electrode interaction that can form distinct virtual channels.
Electrical Impedance Tomography: A Reconstruction Method Based on Neural Networks and Particle Swarm Optimization
(2015-01-01)
Martin, Sebastien
;
Choi, Charles T. M.
;
交大名義發表
;
National Chiao Tung University
Electrical Impedance Tomography (EIT) is a non-invasive image reconstruction technique. Typically, an EIT scheme involves the solution to an inverse problem, which usually gives a poor resolution, due to linearization and ill-posedness of the problem. An alternative approach based on Artificial Neural Networks (ANN) has been used as a replacement of the inverse problem, giving correct results without linearizing the problem. However, training an ANN may be time consuming and usually requires a large amount of iterations before achieving a correct answer to the input stimulation. Several studies focused on training ANNs, and Evolutionary Algorithms (EA) gives a faster global convergence. In this paper, a novel approach based on Artificial Neural Networks and Particle Swarm Optimization (PSO) is proposed to improve the training process. A training method based on PSO algorithm achieves a faster global convergence.
Envelope ADI-FDTD method and its application in three-dimensional nonuniform meshes
(2007-04-01)
Sun, Shu-Hai
;
Choi, Charles T. M.
;
資訊工程學系
;
Department of Computer Science
The envelope alternating-direction-implicit finite difference time domain (ADI-FDTD) method in 3-D nonuniform meshes was proposed and studied. The phase velocity error for the envelope ADI-FDTD and ADI-FDTD methods in uniform and nonuniform meshes and different temporal increments were studied. A cavity problem was studied using the envelope ADI-FDTD and ADI-FDTD methods in graded meshes and the conventional FDTD method in a uniform mesh. The simulation results show that the envelope ADI-FDTD performs better than the ADI-FDTD in numerical accuracy.
Extracting Speech Signals using Independent Component Analysis
(2009)
Choi, Charles T. M.
;
Lee, Yi-Hsuan
;
資訊工程學系
;
Department of Computer Science
Independent component analysis (ICA) is the dominant method to resolve blind source separation (BSS) problem. In this article we conducted experiments to evaluate the separation performance of ICA for acoustic signals. Experiments results show that if we can find appropriate placement of microphones, applying ICA to hearing prostheses as pre-processing can help the wearer hear more clear sounds.
Fast and Accurate Solution of the Inverse Problem for Image Reconstruction Using Electrical Impedance Tomography
(2019-06-01)
Martin, Sebastien
;
Choi, Charles T. M.
;
分子醫學與生物工程研究所
;
電機工程學系
;
電控工程研究所
The goal in electrical impedance tomography is to obtain the electrical properties of different materials by applying an electrical current and measuring the resulting potential difference at the boundaries of the domain. While the numerical accuracy is technically limited by the size of the elements within the finite-element (FE) mesh, using a fine mesh will result in a computationally demanding reconstruction, especially when the region of interest (ROI) is not known. However, this situation is different when the location is known, when one can easily refine the FE model around the target, aiming for greater accuracy around the ROI. In this paper, an innovative approach estimates the location of the target object before solving the inverse problem, so that it becomes possible to refine only a specific area of the FE model. A powerful artificial intelligence method is used to obtain this region.
Finite Element Models of Neuron Electrode Sealing Interfaces
(2012-02-01)
Choi, Charles T. M.
;
You, Shan-Jen
;
分子醫學與生物工程研究所
;
資訊工程學系
;
電機工程學系
It is desirable to detect any leakage current when microelectrode is used to stimulate a neuron electrically. This paper proposes a new approach to study the neuron-electrode sealing interface problem. As opposite to the traditional bi-domain FEM that need a two-step process of indirect coupling of two domains with a circuit equation, which involves solving a set of ordinary differential equation, this paper proposed a more elegant approach to study the neuron-electrode sealing interface problem based on a single domain finite element model. The result shows the stimulation electrical potential distribution and the sealing resistance match the published simulation and experimental results.
Incorporating the electrode-tissue interface to cochlear implant models
(2007-04-01)
Lai, Wei-Dian
;
Choi, Charles T. M.
;
資訊工程學系
;
腦科學研究中心
;
Department of Computer Science
The electrode is an important part of cochlear implant system because it affects the current spread and the response of the auditory nerves. In this paper, a new electrode-tissue interface model is coupled with a 3D cochlea and electrode model to simulate the condition of the electrodes in human cochlea is proposed. The effect of incorporating the electrode-tissue interface in a cochlea and the response of the auditory nerves is discussed. Three models are studied, a model of electrode without interface, a model of electrode incorporating the interface effect through an equivalent circuit, and a new model of electrode incorporating a thin layer between the electrode and the scala typmani. The three electrode models were studied using activating functions. Significant different results were found for electrodes with or without taking interface into account. The model results for the electrode model taking into account of the interface effect using an equivalent circuit and a thin interface layer are almost identical.
Modeling Deep Brain Stimulation
(2009)
Choi, Charles T. M.
;
Lee, Yen-Ting
;
資訊工程學系
;
Department of Computer Science
The goal of this project was to develop a quantitative understanding about shape of axonal tissue activated (STA) and stimulation site by deep brain stimulation (DBS) in middle brain like thalamus, subthalamic nucleus (STN), and globus pollidus interna (GPi) to make better stimulation effects. The volume of tissue activated (VTA) can be studied using different electrode dimensions.
Modeling Deep Brain Stimulation Based on Current Steering Scheme
(2011-05-01)
Choi, Charles T. M.
;
Lee, Yen-Ting
;
Tsou, Yi-Lin
;
分子醫學與生物工程研究所
;
資訊工程學系
The goal of this paper is to develop a quantitative understanding about the shape and the volume of tissue activated (VTA) and stimulation sites by deep brain stimulation (DBS) in deep brain area like subthalamic nucleus (STN) to improve neural stimulation. Monopolar current steering approach has been studied in cochlear implants and deep brain stimulation. In this paper, novel bipolar and tripolar current steering schemes are applied to deep brain stimulation. Finite-element models of deep brain stimulation are used to study the volume of tissue activated for monopolar, bipolar, and tripolar current steering configurations.
Modeling ECAP in Cochlear Implants Using the FEM and Equivalent Circuits
(2014-02-01)
Choi, Charles T. M.
;
Wang, Shao Po
;
分子醫學與生物工程研究所
;
電機資訊學士班
;
Institute of Molecular Medicine and Bioengineering
Evoked compound action potential (ECAP) has been used clinically to determine whether auditory nerves are responding to electrical stimulation in cochlear implant systems. In this paper, a novel scheme is proposed to model ECAP and its measurement using a generalized Schwarz-Eikhof-Frijns nerve model and equivalent circuits based on the finite-element (FE) method. A 3-D FE model of a cochlear implant is used in this scheme. The ECAP modeling result is validated with clinical ECAP measurement results.
Noise Reduction Using Wavelet Thresholding of Multitaper Estimators and Geometric Approach to Spectral Subtraction for Speech Coding Strategy
(2012-04-01)
Chu, Kai Chuan
;
Choi, Charles T. M.
;
分子醫學與生物工程研究所
;
資訊工程學系
;
電機工程學系
Objectives. Noise reduction using wavelet thresholding of multitaper estimators (WTME) and geometric approach to spectral subtraction (GASS) can improve speech quality of noisy sound for speech coding strategy. This study used Perceptual Evaluation of Speech Quality (PESQ) to assess the performance of the WTME and GASS for speech coding strategy. Methods. This study included 25 Mandarin sentences as test materials. Environmental noises including the air-conditioner, cafeteria and multi-talker were artificially added to test materials at signal to noise ratio (SNR) of -5, 0, 5, and 10 dB. HiRes 120 vocoder WTME and GASS noise reduction process were used in this study to generate sound outputs. The sound outputs were measured by the PESQ to evaluate sound quality. Results. Two figures and three tables were used to assess the speech quality of the sound output of the WTME and GASS. Conclusion. There is no significant difference between the overall performance of sound quality in both methods, but the geometric approach to spectral subtraction method is slightly better than the wavelet thresholding of multitaper estimators.
Nonlinear Electrical Impedance Tomography Reconstruction Using Artificial Neural Networks and Particle Swarm Optimization
(2016-03)
Martin, Sebastien
;
Choi, Charles T. M.
;
分子醫學與生物工程研究所
;
電機學院
;
Institute of Molecular Medicine and Bioengineering
Electrical impedance tomography (EIT) is an imaging technology that offers the advantages of being noninvasive, and it does not generate ionizing radiation. The main difficulty in applying EIT is to solve an ill-posed nonlinear inverse problem. Given a set of electrical voltages measured at the surface of a volume conductor, the goal is to identify the materials that are present in the domain by determining their electrical conductivities. However, since EIT is a nonlinear problem, various algorithms proposed in the literature can only approximate real conductivity distributions. Nonlinear algorithms, especially artificial neural networks (ANNs), have been proposed to solve this inverse problem, but these algorithms are usually limited by slow convergence issues during the training phase. In this paper, the particle swarm optimization (PSO) method is used to train an ANN to solve the EIT problem. It has been found that, compared with the back-propagation algorithm, PSO is capable of generating both faster and higher convergence. This paper also shows that the proposed method is capable of dealing with noisy data and the imperfections in the finite-element discretization, an important source of errors in EIT imaging.
On the influence of spread constant in radial basis networks for electrical impedance tomography
(2016-06-01)
Martin, Sebastien
;
Choi, Charles T. M.
;
交大名義發表
;
National Chiao Tung University
Electrical impedance tomography (EIT) is a non-invasive imaging technique. The main task of this work is to solve a non-linear inverse problem, for which several techniques have been suggested, but none of which gives a very high degree of accuracy. This paper introduces a novel approach, based on radial basis function (RBF) artificial neural networks (ANNs), to solve this problem, and uses several ANNs to obtain the best solution to the EIT inverse problem. ANNs have the potential to directly estimate the solution of the inverse problem with a high degree of accuracy. While different radial basis neural networks do not always perform well on different problems, they usually give good results on some specific problems. This paper evidences a strong correlation between the area of the target and the spread constant of the RBF network that gives the best reconstruction. A solution to automatically estimate the size of the target and pick the best neural network directly from voltage measurements is presented, making the reconstruction process automatic. By automatically selecting the best ANN for each specific set of voltage measurements, the proposed solution gives a more accurate reconstruction of both small and large targets.
On the influence of spread constant in radial basis networks for electrical impedance tomography
(2016-06)
Martin, Sebastien
;
Choi, Charles T. M.
;
交大名義發表
;
National Chiao Tung University
Electrical impedance tomography (EIT) is a non-invasive imaging technique. The main task of this work is to solve a non-linear inverse problem, for which several techniques have been suggested, but none of which gives a very high degree of accuracy. This paper introduces a novel approach, based on radial basis function (RBF) artificial neural networks (ANNs), to solve this problem, and uses several ANNs to obtain the best solution to the EIT inverse problem. ANNs have the potential to directly estimate the solution of the inverse problem with a high degree of accuracy. While different radial basis neural networks do not always perform well on different problems, they usually give good results on some specific problems. This paper evidences a strong correlation between the area of the target and the spread constant of the RBF network that gives the best reconstruction. A solution to automatically estimate the size of the target and pick the best neural network directly from voltage measurements is presented, making the reconstruction process automatic. By automatically selecting the best ANN for each specific set of voltage measurements, the proposed solution gives a more accurate reconstruction of both small and large targets.