Noise Correction Techniques Research Articles

W2P: Switching from Weak Supervision to Partial Supervision for Semantic Segmentation

Current weakly-supervised semantic segmentation (WSSS) techniques concentrate on enhancing class activation maps (CAMs) with image-level annotations. Yet, the emphasis on producing these pseudo-labels often overshadows the pivotal role of training the segmentation model itself. This paper underscores the significant influence of noisy pseudo-labels on segmentation network performance, particularly in boundary region. To address above issues, we introduce a novel paradigm: Weak to Partial Supervision (W2P). At its core, W2P categorizes the pseudo-labels from WSSS into two unique supervisions: trustworthy clean labels and uncertain noisy labels. Next, our proposed partially-supervised framework adeptly employs these clean labels to rectify the noisy ones, thereby promoting the continuous enhancement of the segmentation model. To further optimize boundary segmentation, we incorporate a noise detection mechanism that specifically preserves boundary regions while eliminating noise. During the noise refinement phase, we adopt a boundary-conscious noise correction technique to extract comprehensive boundaries from noisy areas. Furthermore, we devise a boundary generation approach that assists in predicting intricate boundary zones. Evaluations on the PASCAL VOC 2012 and MS COCO 2014 datasets confirm our method's impressive segmentation capabilities across various pseudo-labels.

Ensemble fluid simulations on quantum computers

We discuss the viability of ensemble simulations of fluid flows on quantum computers. The basic idea is to formulate a functional Liouville equation for the probability distribution of the flow field configuration and recognize that, due to its linearity, such an equation is in principle more amenable to quantum computing than the dynamic equations of fluid motion. After suitable marginalization and associated closure, the Liouville approach is shown to require several hundreds of logical qubits, hence calling for a major thrust in current noise correction and mitigation techniques.

Pixel-accurate road crack detection in presence of inaccurate annotations

Recent road crack detection methods obtain appealing scores but typically allow a few pixel tolerance margin. This is acceptable for locating cracks, but not for measuring their width (indicator of the cracks’ severity). Our baseline model, U-VGG19, obtains an F-score of 71.77% on CrackForest, which is superior to other approaches when no tolerance is admitted. However, increasing the scores without tolerance is difficult due to inaccurate annotations. We propose a novel synthetic dataset, Syncrack, as a benchmark for the evaluation of training with inaccurate annotations. Our results show that inaccurate annotations have a detrimental impact on the F-measure, decreasing it by up to 20%. To overcome this, we study label noise correction techniques using weakly supervised learning. Training U-VGG19 with these corrected labels improves the results on Syncrack by up to 12%. Obtained results on the CrackForest and Aigle-RN datasets support that these approaches are useful for real-life data too.

Physiological noise modeling in fMRI based on the pulsatile component of photoplethysmograph

The blood oxygenation level-dependent (BOLD) contrast mechanism allows the noninvasive monitoring of changes in deoxyhemoglobin content. As such, it is commonly used in functional magnetic resonance imaging (fMRI) to study brain activity since levels of deoxyhemoglobin are indirectly related to local neuronal activity through neurovascular coupling mechanisms. However, the BOLD signal is severely affected by physiological processes as well as motion. Due to this, several noise correction techniques have been developed to correct for the associated confounds. The present study focuses on cardiac pulsatility fMRI confounds, aiming to refine model-based techniques that utilize the photoplethysmograph (PPG) signal. Specifically, we propose a new technique based on convolution filtering, termed cardiac pulsatility model (CPM) and compare its performance with the cardiac-related RETROICOR (Card-RETROICOR), which is a technique commonly used to model fMRI fluctuations due to cardiac pulsatility. Further, we investigate whether variations in the amplitude of the PPG pulses (PPG-Amp) covary with variations in amplitude of pulse-related fMRI fluctuations, as well as with the systemic low frequency oscillations (SLFOs) component of the fMRI global signal (GS – defined as the mean signal across all gray matter voxels). Capitalizing on 3T fMRI data from the Human Connectome Project, CPM was found to explain a significantly larger fraction of the fMRI signal variance compared to Card-RETROICOR, particularly for subjects with larger heart rate variability during the scan. The amplitude of the fMRI pulse-related fluctuations did not covary with PPG-Amp; however, PPG-Amp explained significant variance in the GS that was not attributed to variations in heart rate or breathing patterns. Our results suggest that the proposed approach can model high-frequency fluctuations due to pulsation as well as low-frequency physiological fluctuations more accurately compared to model-based techniques commonly employed in fMRI studies.

Unravelling the Role of Aging During Cold Rolling on the Primary Recrystallization of Grain-Oriented Electrical Steels Using a Robust EBSD-Based Analysis Approach

Aging during cold rolling improves the final magnetic properties of conventionally processed grain oriented electrical steels (GOES). In order to determine the influence of aging on the evolution of microstructure and texture during recrystallization, identical hot rolled GOES sheets, were cold rolled with and without aging, and examined by large-area electron backscatter diffraction (EBSD) investigations. Local misorientations were determined from the EBSD maps with a noise-correction technique and the statistical relevance of the EBSD-based texture measurements was examined and confirmed by a new approach. The most important effect of aging was found to be a significant modification of the deformation and recrystallization behaviour of the α-fiber components. In fact, the aging treatment resulted in more shear banding during cold rolling, which is attributed to dislocation pinning by carbon atoms. The dislocation pinning furthermore leads to significantly reduced recovery during annealing, which finally results in a higher remaining driving force for recrystallization and more nucleation sites in the otherwise sluggishly recrystallizing α-fiber grains. As a result the grain size of the aged material is significantly reduced. We assume that this smaller grain size provides a higher driving force for the subsequent growth of the Goss grains and leads to improved magnetic properties of the final processed GOES, although these relations have not been investigated here. In contrast, the α*-fiber, which is often suggested to be of importance for optimized Goss texture formation, was not found to be affected significantly by the aging treatment.

An Improved Rician Noise Correction Technique from the Magnitude of Diffusion MR Images

An Improved Rician Noise Correction Technique from the Magnitude of Diffusion MR Images

Characterisation of Tidal Flows at the European Marine Energy Centre in the Absence of Ocean Waves

The data analyses and results presented here are based on the field measurement campaign of the Reliable Data Acquisition Platform for Tidal (ReDAPT) project (Energy Technologies Institute (ETI), U.K. 2010–2015). During ReDAPT, a 1 MW commercial prototype tidal turbine was deployed and operated at the Fall of Warness tidal test site within the European Marine Energy Centre (EMEC), Orkney, U.K. Mean flow speeds and Turbulence Intensity (TI) at multiple positions proximal to the machine are considered. Through the implemented wave identification techniques, the dataset can be filtered into conditions where the effects of waves are present or absent. Due to the volume of results, only flow conditions in the absence of waves are reported here. The analysis shows that TI and mean flows are found to vary considerably between flood and ebb tides whilst exhibiting sensitivity to the tidal phase and to the specification of spatial averaging and velocity binning. The principal measurement technique was acoustic Doppler profiling provided by seabed-mounted Diverging-beam Acoustic Doppler Profilers (D-ADP) together with remotely-operable Single-Beam Acoustic Doppler Profilers (SB-ADP) installed at mid-depth on the tidal turbine. This novel configuration allows inter-instrument comparisons, which were conducted. Turbulence intensity averaged over the rotor extents of the ReDAPT turbine for flood tides vary between 16.7% at flow speeds above 0.3 m/s and 11.7% when considering only flow speeds in the turbine operating speed range, which reduces to 10.9% (6.8% relative reduction) following the implementation of noise correction techniques. Equivalent values for ebb tides are 14.7%, 10.1% and 9.3% (7.9% relative reduction). For flood and ebb tides, TI values resulting from noise correction are reduced in absolute terms by 3% and 2% respectively across a wide velocity range and approximately 1% for turbine operating speeds. Through comparison with SB-ADP-derived mid-depth TI values, this correction is shown to be conservative since uncorrected SB-ADP results remain, in relative terms, between 10% and 21% below corrected D-ADP values depending on tidal direction and the range of velocities considered. Results derived from other regions of the water column, those important to floating turbine devices for example, are reported for comparison.

Improving Crowdsourced Label Quality Using Noise Correction.

Crowdsourcing systems provide a cost effective and convenient way to collect labels, but they often fail to guarantee the quality of the labels. This paper proposes a novel framework that introduces noise correction techniques to further improve the quality of integrated labels that are inferred from the multiple noisy labels of objects. In the proposed general framework, information about the qualities of labelers estimated by a front-end ground truth inference algorithm is utilized to supervise subsequent label noise filtering and correction. The framework uses a novel algorithm termed adaptive voting noise correction (AVNC) to precisely identify and correct the potential noisy labels. After filtering out the instances with noisy labels, the remaining cleansed data set is used to create multiple weak classifiers, based on which a powerful ensemble classifier is induced to correct these noises. Experimental results on eight simulated data sets with different kinds of features and two real-world crowdsourcing data sets in different domains consistently show that: 1) the proposed framework can improve label quality regardless of inference algorithms, especially under the circumstance that each instance has a few repeated labels and 2) since the proposed AVNC algorithm considers both the number of and the probability of potential label noises, it outperforms the state-of-the-art noise correction algorithms.

Test-retest reliability of the default mode network in a multi-centric fMRI study of healthy elderly: Effects of data-driven physiological noise correction techniques.

Understanding how to reduce the influence of physiological noise in resting state fMRI data is important for the interpretation of functional brain connectivity. Limited data is currently available to assess the performance of physiological noise correction techniques, in particular when evaluating longitudinal changes in the default mode network (DMN) of healthy elderly participants. In this 3T harmonized multisite fMRI study, we investigated how different retrospective physiological noise correction (rPNC) methods influence the within-site test-retest reliability and the across-site reproducibility consistency of DMN-derived measurements across 13 MRI sites. Elderly participants were scanned twice at least a week apart (five participants per site). The rPNC methods were: none (NPC), Tissue-based regression, PESTICA and FSL-FIX. The DMN at the single subject level was robustly identified using ICA methods in all rPNC conditions. The methods significantly affected the mean z-scores and, albeit less markedly, the cluster-size in the DMN; in particular, FSL-FIX tended to increase the DMN z-scores compared to others. Within-site test-retest reliability was consistent across sites, with no differences across rPNC methods. The absolute percent errors were in the range of 5-11% for DMN z-scores and cluster-size reliability. DMN pattern overlap was in the range 60-65%. In particular, no rPNC method showed a significant reliability improvement relative to NPC. However, FSL-FIX and Tissue-based physiological correction methods showed both similar and significant improvements of reproducibility consistency across the consortium (ICC = 0.67) for the DMN z-scores relative to NPC. Overall these findings support the use of rPNC methods like tissue-based or FSL-FIX to characterize multisite longitudinal changes of intrinsic functional connectivity. Hum Brain Mapp 37:2114-2132, 2016. © 2016 Wiley Periodicals, Inc.

Metaheuristic Based Noise Identification and Image Denoising Using Adaptive Block Selection Based Filtering

Image denoising has become one of the major forms of image enhancement methods that form the basis of image processing. Due to the inconsistencies in the machinery producing these signals, medical images tend to require these techniques. In real time, images do not contain a single noise, and instead they contain multiple types of noise distributions in several indistinct regions. This paper presents an image denoising method that uses Metaheuristics to perform noise identification. Adaptive block selection is used to identify and correct the noise contained in these blocks. Though the system uses a block selection scheme, modifications are performed on pixel- to-pixel basis and not on the entire blocks; hence the image accuracy is preserved. PSO is used to identify the noise distribution, and appropriate noise correction techniques are applied to denoise the images. Experiments were conducted using salt and pepper noise, Gaussian noise and a combination of both the noise in the same image. It was observed that the proposed method performed effectively on noise levels up-to 0.5 and was able to produce results with PSNR values ranging from 20 to 30 in most of the cases. Excellent reduction rates were observed on salt and pepper noise and moderate reduction rates were observed on Gaussian noise. Experimental results show that our proposed system has a wide range of applicability in any domain specific image denoising scenario, such as medical imaging, mammogram etc.

Physiological brainstem mechanisms of trigeminal nociception: An fMRI study at 3T

The brainstem is a major site of processing and modulation of nociceptive input and plays a key role in the pathophysiology of various headache disorders. However, human imaging studies on brainstem function following trigeminal nociceptive stimulation are scarce as brainstem specific imaging approaches have to address multiple challenges such as magnetic field inhomogeneities and an enhanced level of physiological noise. In this study we used a viable protocol for brainstem fMRI of standardized trigeminal nociceptive stimulation to achieve detailed insight into physiological brainstem mechanisms of trigeminal nociception. We conducted a study of 21 healthy participants using a nociceptive ammonia stimulation of the left nasal mucosa with an optimized MR acquisition protocol for high resolution brainstem echoplanar imaging in combination with two different noise correction techniques. Significant BOLD responses to noxious ammonia stimulation were observed in areas typically involved in trigeminal nociceptive processing such as the spinal trigeminal nuclei (sTN), thalamus, secondary somatosensory cortex, insular cortex and cerebellum as well as in a pain modulating network including the periaqueductal gray area, hypothalamus (HT), locus coeruleus and cuneiform nucleus (CNF). Activations of the left CNF were positively correlated with pain intensity ratings. Employing psychophysiological interaction (PPI) analysis we found enhanced functional connectivity of the sTN with the contralateral sTN and HT following trigeminal nociception. We also observed enhanced functional connectivity of the CNF with the RVM during painful stimulation thus implying an important role of these two brainstem regions in central pain processing. The chosen approach to study trigeminal nociception with high-resolution fMRI offers new insight into human pain processing and might thus lead to a better understanding of headache pathophysiology.

An Improved De-Noising Algorithm for Highly Corrupted Color Videos Using FPGA Based Impulse Noise Detection and Correction Techniques

Objective : This research focuses on design and implementation of FPGA hardware architecture based image de-noising algorithm with automatic detection and correction of impulse noise. With the proposed hardware, image or video camera can be interfaced with the implementing hardware for de-noising of salt and pepper noise called impulse noise. Methods/Analysis : The algorithms proposed in this research mainly work on images to identify the impulse noise affected pixel and correct only the corrupted pixel instead of uncorrupted. Also the novelty method modifies the existed features of various de-noising techniques using real time, low power FPGA architecture for the noise detection and correction. Findings : The proposed research is in two stages namely modified boundary discriminative noise detection and recursive median filter based correction technique. These two stages are tested for various noise densities and delivered better de-noising factor than the existing algorithms. The simulation results of the implemented algorithm prove efficiency of 98% noise reduction factor for 90% corrupted image or video.

The influence of physiological noise correction on test-retest reliability of resting-state functional connectivity.

The utility and success of resting-state functional connectivity MRI (rs-fcMRI) depend critically on the reliability of this technique and the extent to which it accurately reflects neuronal function. One challenge is that rs-fcMRI is influenced by various sources of noise, particularly cardiac- and respiratory-related signal variations. The goal of the current study was to evaluate the impact of various physiological noise correction techniques, specifically those that use independent cardiac and respiration measures, on the test-retest reliability of rs-fcMRI. A group of 25 subjects were each scanned at three time points--two within the same imaging session and another 2-3 months later. Physiological noise corrections accounted for significant variance, particularly in blood vessels, sagittal sinus, cerebrospinal fluid, and gray matter. The fraction of variance explained by each of these corrections was highly similar within subjects between sessions, but variable between subjects. Physiological corrections generally reduced intrasubject (between-session) variability, but also significantly reduced intersubject variability, and thus reduced the test-retest reliability of estimating individual differences in functional connectivity. However, based on known nonneuronal mechanisms by which cardiac pulsation and respiration can lead to MRI signal changes, and the observation that the physiological noise itself is highly stable within individuals, removal of this noise will likely increase the validity of measured connectivity differences. Furthermore, removal of these fluctuations will lead to better estimates of average or group maps of connectivity. It is therefore recommended that studies apply physiological noise corrections but also be mindful of potential correlations with measures of interest.

Task-related BOLD responses and resting-state functional connectivity during physiological clamping of end-tidal CO2

Carbon dioxide (CO(2)), a potent vasodilator, is known to have a significant impact on the blood-oxygen level dependent (BOLD) signal. With the growing interest in studying synchronized BOLD fluctuations during the resting state, the extent to which the apparent synchrony is due to variations in the end-tidal pressure of CO(2) (PETCO(2)) is an important consideration. CO(2)-related fluctuations in BOLD signal may also represent a potential confound when studying task-related responses, especially if breathing depth and rate are affected by the task. While previous studies of the above issues have explored retrospective correction of BOLD fluctuations related to arterial PCO(2), here we demonstrate an alternative approach based on physiological clamping of the arterial CO(2) level to a near-constant value. We present data comparing resting-state functional connectivity within the default-mode-network (DMN), as well as task-related BOLD responses, acquired in two conditions in each subject: 1) while subject's PETCO(2) was allowed to vary spontaneously; and 2) while controlling subject's PETCO(2) within a narrow range. Strong task-related responses and areas of maximal signal correlation in the DMN were not significantly altered by suppressing fluctuations in PETCO(2). Controlling PETCO(2) did, however, improve the performance of retrospective physiological noise correction techniques, allowing detection of additional regions of task-related response and resting-state connectivity in highly vascularized regions such as occipital cortex. While these results serve to further rule out systemic physiological fluctuations as a significant source of apparent resting-state network connectivity, they also demonstrate that fluctuations in arterial CO(2) are one of the factors limiting sensitivity in task-based and resting-state fMRI, particularly in regions of high vascular density. This must be considered when comparing subject groups who might exhibit differences in respiratory physiology or breathing patterns.

Noise-signal index threshold: a new noise-reduction technique for generation of reference spectra and efficient hyperspectral image classification

Reference spectra of terrestrial targets are usually collected using field spectro-radiometers for mineral abundance mapping and target detection. These spectra often have noise that masks characteristic absorption and reflection features and affects the efficiency of material mapping. This work aims at obtaining an empirical technique for reduction of high-frequency noise from field spectra. The proposed noise correction technique uses a ‘normalized’ measure Rn , where Rn = (Ln − Fn )/Ln for each band (n) calculated from field and laboratory spectra of test material, with Fn and Ln being the depth of the absorption feature in field and laboratory spectra, respectively. On the basis of the assumption of the constancy of this ratio in neighbouring bands, an empirical algorithm that approximates the ratio Rn of a noisy band to the corrected ratio of an adjacent band is used to obtain the noise-corrected field spectra. The classification accuracy increases significantly when noise reduced field spectra are used as reference spectra.

The non-separability of physiologic noise in functional connectivity MRI with spatial ICA at 3 T

The impact of physiologic noise on spatial ICA analyses of resting state BOLD-weighted MRI data is investigated. Using FastICA and Infomax ICA, two common ICA algorithms, we apply a group spatial ICA method across multiple subjects. We compare the spatial maps from five commonly identified functional networks and show that physiologic noise correction techniques introduce significant changes in the spatial ICA decomposition of all five networks, greater than the changes introduced by either algorithmic indeterminacy (re-running ICA) or the changes introduced by decreasing the decomposition dimensionality due to physiologic noise removal. In addition, we demonstrate that the sources associated with these components have significant temporal correlation to parallel measures of cardiac and respiratory rates, and these are reduced after correction. We conclude that ICA decomposition is significantly affected by physiologic noise and the ICA process alone is not sufficient to separate physiologic noise effects in the brain. It is recommended that physiologic noise correction be applied to timeseries data prior to ICA decomposition.

Poster - Thurs Eve-28: New brain diffusion analysis method: White matter grey matter dissasociation.

Diffusion MR studies are often used to investigate the physical properties of brain tissues (1, 2). It is known that a full characterisation of the diffusion decay for brain could give valuable information about the structural organisation of cerebral tissue. The significance of the present diffusion decay study lies in the combination of three novel procedures to provide a better characterization of the diffusion decay: i) the acquisition of a large number of b-values (96 b-values up to 10,000 s/mm2 ), ii) the application of a noise correction technique (3) to the acquired data, and iii) the use of a Non Negative Least Squares (NNLS) fitting algorithm to evaluate the diffusion coefficients. The presence of noise in magnitude MR images can affect the calculation of the diffusion parameters (4) and therefore a noise correction technique (3) is applied. The NNLS algorithm is used to fit the corrected data instead of the more commonly used Levenberg-Marquardt algorithm since the NNLS algorithm does not require the number of components to be specified, nor does it need initial estimates of the fitting parameters as input; thus giving it more versatility as a fitting tool for the diffusion decay. The results indicate that the diffusion decays in grey and white matter have one and two components, respectively. Consequently, the short diffusion component in white matter (Fig. 1.c) can be used as a tool in the disassociation of white and grey matter tissues.

Integration of motion correction and physiological noise regression in fMRI

Physiological fluctuations resulting from the heart beat and respiration are a dominant source of noise in fMRI, particularly at high field strengths. Commonly used physiological noise correction techniques, such as RETROspective Image CORrection (RETROICOR), rely critically on the timing of the image acquisition relative to the heart beat, but do not account for the effects of subject motion. Such motion affects the fluctuation amplitude, yet volume registration can distort the timing information. In this study, we aimed to systematically determine the optimal order of volume registration, slice-time correction and RETROICOR in their traditional forms. In addition, we evaluate the sensitivity of RETROICOR to timing errors introduced by the slice acquisition, and we develop a new method of accounting for timing errors introduced by volume registration into physiological correction (motion-modified RETROICOR). Both simulation and resting data indicate that the temporal standard deviation is reduced most by performing volume registration before RETROICOR and slice-time correction after RETROCIOR. While simulations indicate that physiological noise correction with regressors constructed on a slice-by-slice basis more accurately modeled physiological noise compared to using the same regressors for the entire volume, the difference between these regression techniques in subject data was minimal. The motion-modified RETROICOR showed marked improvement in simulations with varying amounts of subject motion, reducing the temporal standard deviation by up to 36% over the traditional RETROICOR. Though to a lesser degree than in simulation, the motion-modified RETROICOR performed better in nearly every voxel in the brain in both high- and low-resolution subject data.

Quantification of resolution and noise effects on thermal dissipation measurements in turbulent non-premixed jet flames

One-dimensional (1-D) line Rayleigh thermometry is used to investigate the effects of spatial resolution and noise on thermal dissipation in turbulent non-premixed CH 4/H 2/N 2 jet flames. The high signal-to-noise ratio and spatial resolution of the measured temperature field enables determination of the cutoff wavenumber in the 1-D temperature dissipation spectrum obtained at each flame location. The local scale inferred from this cutoff is analogous to the Batchelor scale in nonreacting flows. At downstream locations in the flames studied here, it is consistent with estimates of the Batchelor scale based on the scaling laws using local Reynolds numbers. The spectral cutoff information is used to design data analysis schemes for determining mean thermal dissipation. Laminar flame measurements are used to characterize experimental noise and correct for the noise-induced apparent dissipation in the turbulent flame results. These experimentally determined resolution and noise correction techniques are combined to give measurements of the mean thermal dissipation that are essentially fully resolved and noise-free. The prospects of using spectral results from high-resolution 1-D Rayleigh imaging measurements to design filtering schemes for Raman-based measurements of mixture fraction dissipation are also discussed.

Postroute Gate Sizing for Crosstalk Noise Reduction

Gate sizing is a practical and a feasible crosstalk noise correction technique in the post route design stage, especially for block level sea-of-gates designs. The difficulty in gate sizing for noise reduction is that, by increasing a driver size, noise at the driver output is reduced, but noise injected by that driver on other nets is increased. This can create cyclical dependencies between nets in the circuit with noise violations. In this paper, we propose a fast and effective heuristic postroute gate-sizing algorithm that uses a graph representation of the noise dependencies between nodes. Our method utilizes gate sizing in both directions and works in linear time as a function of the number of gates. The effectiveness of the algorithm is shown on several industrial high-performance designs.