Showing papers by "Jari P. Kaipio published in 2017"
TL;DR: In this article, a truncated Fourier series approximation approach was proposed to speed up the computation of thermal tomography by solving a frequency domain equivalent of the time domain heat diffusion equation at multiple frequencies.
13 citations
10 citations
8 citations
TL;DR: The authors' GRIN retrieval algorithm produces fast and accurate measurements of the lens GRIN profile, and experiments to study the optics of physiologically perturbed lenses are the future direction of this research.
Abstract: Measuring the lens gradient refractive index (GRIN) accurately and reliably has proven an extremely challenging technical problem. A fully automated laser ray tracing (LRT) system was built to address this issue. The LRT system captures images of multiple laser projections before and after traversing through an ex vivo lens. These LRT images, combined with accurate measurements of the lens geometry, are used to calculate the lens GRIN profile. Mathematically, this is an ill-conditioned problem; hence, it is essential to apply biologically relevant constraints to produce a feasible solution. The lens GRIN measurements were compared with previously published data. Our GRIN retrieval algorithm produces fast and accurate measurements of the lens GRIN profile. Experiments to study the optics of physiologically perturbed lenses are the future direction of this research.
8 citations
TL;DR: A statistical method based on the Bayesian approximation error approach to compensate for source imaging errors related to erronous skull conductivity is proposed and demonstrated by simulating EEG data of focal source activity and using the dipole scan algorithm and a sparsity promoting prior to reconstruct the underlying sources.
Abstract: Knowing the correct skull conductivity is crucial for the accuracy of EEG source imaging, but unfortunately, its true value, which is inter- and intra-individually varying, is difficult to determine. In this paper, we propose a statistical method based on the Bayesian approximation error approach to compensate for source imaging errors related to erronous skull conductivity. We demonstrate the potential of the approach by simulating EEG data of focal source activity and using the dipole scan algorithm and a sparsity promoting prior to reconstruct the underlying sources. The results suggest that the greatest improvements with the proposed method can be achieved when the focal sources are close to the skull.
4 citations
03 Sep 2017
TL;DR: In this paper, the poroelastic signature from an aquifer is simulated and using this signature, the authors estimate aquifer dimensions and hydrologic parameters using seismic data.
Abstract: This study aims at developing computational tools to estimate aquifer dimensions and hydrologic parameters using seismic data. The poroelastic signature from an aquifer is simulated and using this ...
3 citations
TL;DR: In this paper, the authors find weighting factors within a Bayesian framework for the used ''ell 1/\ell 2'' sparsity prior that the resulting maximum a posterior (MAP) estimates do not favor any particular source location.
Abstract: In electroencephalography (EEG) source imaging, the inverse source estimates are depth biased in such a way that their maxima are often close to the sensors. This depth bias can be quantified by inspecting the statistics (mean and covariance) of these estimates. In this paper, we find weighting factors within a Bayesian framework for the used \(\ell _1/\ell _2\) sparsity prior that the resulting maximum a posterior (MAP) estimates do not favour any particular source location. Due to the lack of an analytical expression for the MAP estimate when this sparsity prior is used, we solve the weights indirectly. First, we calculate the Gaussian prior variances that lead to depth un-biased maximum a posterior (MAP) estimates. Subsequently, we approximate the corresponding weight factors in the sparsity prior based on the solved Gaussian prior variances. Finally, we reconstruct focal source configurations using the sparsity prior with the proposed weights and two other commonly used choices of weights that can be found in literature.
2 citations
TL;DR: In this article, an approach for estimating the optical absorption and scattering directly from the acoustical time series is investigated with simulations, which combines a homogeneous acoustic forward model, based on the Green's function solution of the wave equation, and a finite element method based diffusion approximation model of light propagation into a single forward model.
Abstract: Quantitative photoacoustic tomography seeks to estimate the optical parameters of a target given photoacoustic measurements as a data. Conventionally the problem is split into two steps: 1) the acoustical inverse problem of estimating the acoustic initial pressure distribution from the acoustical time series data; 2) the optical inverse problem of estimating the optical absorption and scattering from the initial pressure distributions. In this work, an approach for estimating the optical absorption and scattering directly from the acoustical time series is investigated with simulations. The work combines a homogeneous acoustical forward model, based on the Green's function solution of the wave equation, and a finite element method based diffusion approximation model of light propagation into a single forward model. This model maps the optical parameters of interest into a time domain signal. The model is used with a Bayesian approach to ill-posed inverse problems to form estimates of the posterior distributions for the parameters of interest. In addition to being able to provide point estimates of the parameters of interest, i.e. reconstruct the absorption and scattering distributions, the approach can be used to derive information on the uncertainty associated with the estimates.
2 citations
01 May 2017
01 Dec 2017
TL;DR: This paper develops a novel approach to adaptive active noise control based on the theory of Bayesian estimation that can be derived as a special case of the well- known FxLMS algorithm, where the noise to be canceled is a white Gaussian process.
Abstract: This paper develops a novel approach to adaptive active noise control based on the theory of Bayesian estimation. Control system parameters are considered as statistical variables and a formulation for the joint probability density function of them is derived. An optimal solution for the system parameters is then calculated through maximizing the density function. An efficient adaptive algorithm for iterative calculation of the optimal parameters is proposed. It is shown that the well- known FxLMS algorithm can be derived as a special case of the proposed algorithm, where the noise to be canceled is a white Gaussian process. Simulation results verify the preference of the proposed system to the traditional active noise control systems in terms of steady-state performance and convergence rate. It is also shown that the preference of the proposed system is much more evident when the noise to be canceled is not white. Finally, a successful implementation of the proposed system in an experimental acoustic duct is reported.