Sparse coding for improved signal-to-noise ratio in MRI

Magnetic Resonance images (MRI) do not only exhibit sparsity but their sparsity take a certain predictable shape which is common for all kinds of images. That region based localised sparsity can be used to de-noise MR images from random thermal noise. This paper present a simple framework to exploit sparsity of MR images for image de-noising. As, noise in MR images tends to change its shape based on contrast level and signal itself, the proposed method is independent of noise shape and type and it can be used in combination with other methods.

Transfer learning using the online Fuzzy Min-Max neural network

In this paper, we present an empirical analysis on transfer learning using the Fuzzy Min–Max (FMM) neural network with an online learning strategy. Three transfer learning benchmark data sets, i.e., 20 Newsgroups, WiFi Time, and Botswana, are used for evaluation. In addition, the data samples are corrupted with white Gaussian noise up to 50 %, in order to assess the robustness of the online FMM network in handling noisy transfer learning tasks. The results are analyzed and compared with those from other methods. The outcomes indicate that the online FMM network is effective for undertaking transfer learning tasks in noisy environments.

Improved image recovery from compressed data contaminated with impulsive noise

Compressed sensing (CS) is a new information sampling theory for acquiring sparse or compressible data with much fewer measurements than those otherwise required by the Nyquist/Shannon counterpart. This is particularly important for some imaging applications such as magnetic resonance imaging or in astronomy. However, in the existing CS formulation, the use of the â„“ 2 norm on the residuals is not particularly efficient when the noise is impulsive. This could lead to an increase in the upper bound of the recovery error. To address this problem, we consider a robust formulation for CS to suppress outliers in the residuals. We propose an iterative algorithm for solving the robust CS problem that exploits the power of existing CS solvers. We also show that the upper bound on the recovery error in the case of non-Gaussian noise is reduced and then demonstrate the efficacy of the method through numerical studies.

Changing listening frequency to minimise white noise and hear indigenous voices

Listening… can involve the listener in an intense, efficacious, and complex set of communicative acts in which one is not speaking, discussing, or disclosing, but sitting quietly, watching, and feeling-the-place, through all the senses…. In the process, one becomes a part of the scene, hearing and feeling with it (Carbaugh 1999: 259).To listen this way involves much more than providing a chance for words to be spoken; it includes tuning in and getting the listening frequency clear. As a non-Indigenous person seeking to conduct qualitative research that listens to Aboriginal people, I need to ask how I can tune into the “active attentiveness” described by Carbaugh (1999) in order to listen in a manner that is appropriate, respectful and minimises my inherent white privilege. In addressing this question I draw on the work of Indigenous authors and academics, critical whiteness studies and my own experiences learning from Aboriginal people in a number of contexts over the past ten to fifteen years.History in Australia since colonization has created a situation where Aboriginal voices are white noise to the ears of many non-Indigenous people. This paper proposes that white privilege and the resulting white noise can be minimised and greater clarity given to Aboriginal voices by privileging Indigenous knowledge and ways of working when addressing Indigenous issues. To minimise the interference of white noise, non-Indigenous people would do well to adopt a position that recognises, acknowledges and utilises some of the strengths that can be learned from Aboriginal culture and Indigenous authors.This paper outlines a model of apprentice, allied listening for non-Indigenous researchers to adopt when preparing to conduct research alongside Indigenous people. Such an approach involves Re-learning of history, Reviewing of the researcher’s beliefs and placing Relating at the centre of the listening approach. Each of these aspects of listening is based on privileging of Indigenous voices.

Faster than the speed of white (noise): potential pressure points in research partnerships and processes

White noise occurs in the thinking, decision making and communication of dominant Settler cultures. It inhibits clear reception of messages, somewhat like the indistinct, fuzzy static of an un-tuned radio. As much a systemic issue as an individual one, it results from assumed privilege and lack of knowledge of worldviews other than the dominant. Until white noise is acknowledged, development of partnerships between Indigenous and non-Indigenous groups is likely to be limited by having to continually start at a point of inequality where nonIndigenous gaps in knowledge and understanding remain unrecognised. This paper/workshop considers challenges encountered while researching experiences of Aboriginal education in Western Australian prisons. Each pressure point occurred where the dominant world view prevailed without question. Discussion will focus on the specific pressure points of ethics approval, project development, informed consent and application of outcomes and findings. The paper asks the questions ‘Who decides what stories are created at these pressure points? What informs those stories?’ As individuals, we might not be able to crash through the white noise barrier but we can chip away and be transparent about its existence with the goal of eventually moving faster than the speed of white (noise).

Robust filtering for uncertain discrete-time systems with uncertain noise covariance and uncertain observations

The use of Kalman filtering is very common in state estimation problems. The problem with Kalman filters is that they require full prior knowledge about the system modeling. It is also assumed that all the observations are fully received. In real applications, the previous assumptions are not true all the time. It is hard to obtain the exact system model and the observations may be lost due to communication problems. In this paper, we consider the design of a robust Kalman filter for systems subject to uncertainties in the state and white noise covariances. The systems under consideration suffer from random interruptions in the measurements process. An upper bound for the estimation error covariance is proposed. The proposed upper bound is further minimized by selection of optimal filter parameters. Simulation example shows the effectiveness of the proposed filter.

Maximum likelihood estimation for rician distributed data in analytical q-ball imaging

Analytical q-ball imaging is widely used for reconstruction of orientation distribution function (ODF) using diffusion weighted MRI data. Estimating the spherical harmonic coefficients is a critical step in this method. Least squares (LS) is widely used for this purpose assuming the noise to be additive Gaussian. However, Rician noise is considered as a more appropriate model to describe noise in MR signal. Therefore, the current estimation techniques are valid only for high SNRs with Gaussian distribution approximating the Rician distribution. The aim of this study is to present an estimation approach considering the actual distribution of the data to provide reliable results particularly for the case of low SNR values. Maximum likelihood (ML) is investigated as a more effective estimation method. However, no closed form estimator is presented as the estimator becomes nonlinear for the noise assumption of the Rician distribution. Consequently, the results of LS estimator is used as an initial guess and the more refined answer is achieved using iterative numerical methods. According to the results, the ODFs reconstructed from low SNR data are in close agreement with ODFs reconstructed from high SNRs when Rician distribution is considered. Also, the error between the estimated and actual fiber orientations was compared using ML and LS estimator. In low SNRs, ML estimator achieves less error compared to the LS estimator.

Application of power functions to chromatographic data for the enhancement of signal to noise ratios and separation resolution

Chromatographic detection responses are recorded digitally. A peak is represented ideally by a Guassian distribution. Raising a Guassian distribution to the power ‘n’ increases the height of the peak to that power, but decreases the standard deviation by √n. Hence there is an increasing disparity in detection responses as the signal moves from low level noise, with a corresponding decrease in peak width. This increases the S/N ratio and increases peak to peak resolution. The ramifications of these factors are that poor resolution in complex chromatographic data can be improved, and low signal responses embedded at near noise levels can be enhanced. The application of this data treatment process is potentially very useful in 2D-HPLC where sample dilution occurs between dimension, reducing signal response, and in the application of post-reaction detection methods, where band broadening is increased by virtue of reaction coils. In this work power functions applied to chromatographic data are discussed in the context of (a) complex separation problems, (b) 2D-HPLC separations, and (c) post-column reaction detectors.