Blind deconvolution: errors, errors everywhere
نویسنده
چکیده
attempt to reconstruct a true spectrum from an observed one. The problem we're considering is sometimes called blind deconvolution, because we're trying to unravel not only the spectrum, but the function that caused the blurring. These problems also arise in image deblurring. Spectroscopy Consider the data in Figure 1, which represents counts measured by a spectrometer. Suppose we have particles whose energy ranges from e lo to e high and define some intermediate energy levels e lo = e 0 < e 1 < … < en b–1 < e n b = e high. This creates n b bins, where the count for the jth bin is the number of particles determined to have energies between ej –1 and ej. Our spectrometer records n b counts, one for each bin; in the figure, we've passed a curve through these counts. Ideally, the count in bin j is exactly the number of particles with energies in the range [e j–1 , e j ]. But some blurring occurs due to the measurement process, and a particle in that energy range might instead be included in the count for a different nearby bin. The probability that a particle with energy e is assigned to bin j is often mod-eled as a normal distribution with mean (e j + e j–1)/2 and variance s j 2. One way to model this system is to try to determine the correct counts f j and the correct blurring given the measured counts g j , j = 1, …, n b and estimates of the values s j. This gives us a matrix equation (K + E) f » g + r, where E accounts for errors in modeling the spectrometer's blur, and r accounts for errors in counts. The matrix entry k jᐉ is computed as the probability that a particle whose energy is in the interval [e ᐉ–1 , e ᐉ ] is assigned to bin j (j, ᐉ = 1, …, n b). There are several sources of differences between the true spectrum and the recorded spectrum: • We effectively assign energy (e j + e j–1)/2 to all particles in bin j, which isn't correct. • A count's value depends on the number of particles with the energies that it represents, but there is some smearing , so it also depends on the number of particles with nearby energies. • …
منابع مشابه
Blind Deconvolution Using a Regularized Structured Total Least Norm Algorithm
Rosen, Park, and Glick proposed the structured total least norm (STLN) algorithm for solving problems in which both the matrix and the right-hand side contain errors. We extend this algorithm for ill-posed problems by adding regularization, and we use the resulting algorithm to solve blind deconvolution problems as encountered in image deblurring when both the image and the blurring function ha...
متن کاملA globally convergent approach for blind MIMO adaptive deconvolution
We discuss the blind deconvolution of multiple input/multiple output (MIMO) linear convolutional mixtures and propose a set of hierarchical criteria motivated by the maximum entropy principle. The proposed criteria are based on the constant–modulus (CM) criterion in order to guarantee that all minima achieve perfectly restoration of different sources. The approach is moreover robust to errors i...
متن کاملConvolutive blind source separation by efficient blind deconvolution and minimal filter distortion
onvolutive blind source separation (BSS) usually encounters two difficulties – the filter indeterminacy in the recovered sources and the relatively high computational load. In this paper we propose an efficient method to convolutive BSS, by dealing with these two issues. It consists of two stages, namely, multichannel blind deconvolution (MBD) and learning the postfilters with the minimum filte...
متن کاملEstimation of input function from dynamic PET brain data using Bayesian blind source separation
Selection of regions of interest in an image sequence is a typical prerequisite step for estimation of time-activity curves in dynamic positron emission tomography (PET). This procedure is done manually by a human operator and therefore suffers from subjective errors. Another such problem is to estimate the input function. It can be measured from arterial blood or it can be searched for a vascu...
متن کاملSimultaneous total variation image inpainting and blind deconvolution
We propose a total variation based model for simultaneous image inpainting and blind deconvolution. We demonstrate that the tasks are inherently coupled together and that solving them individually will lead to poor results. The main advantages of our model are that (i) boundary conditions for deconvolution required near the interface between observed and occluded regions are naturally generated...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
برای دانلود متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید
ثبت ناماگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید
ورودعنوان ژورنال:
- Computing in Science and Engineering
دوره 7 شماره
صفحات -
تاریخ انتشار 2005