Low Signal-to-noise Ratio Research Articles

Mitigating Black-Box Adversarial Attacks via Output Noise Perturbation

In black-box adversarial attacks, attackers query the deep neural network (DNN) and use the query results to optimize the adversarial samples iteratively. In this paper, we study the method of adding white noise to the DNN output to mitigate such attacks. One of our unique contributions is a theoretical analysis of gradient signal-to-noise ratio (SNR), which shows the trade-off between the defense noise level and the attack query cost. The attacker’s query count (QC) is derived mathematically as a function of noise standard deviation. This will guide the defender to find the appropriate noise level for mitigating attacks to the desired security level specified by QC and DNN performance loss. Our analysis shows that the added noise is drastically magnified by the small variation of DNN outputs, which makes the reconstructed gradient have an extremely low SNR. Adding slight white noise with a very small standard deviation, e.g., less than 0.01, is enough to increase QC by many orders of magnitude yet without introducing any noticeable classification accuracy reduction. Our experiments demonstrate that this method can effectively mitigate both soft-label and hard-label black-box attacks under realistic QC constraints. We also prove that this method outperforms many other defense methods and is robust to the attacker’s countermeasures.

Design and Testbed Implementation of Multiuser CFOs Estimation for MIMO SC-FDMA System

In this paper, we propose and implement blind multiuser carrier frequency offsets (CFOs) estimation for multi-input multi-output (MIMO) single carrier frequency division multiple access (SC-FDMA) uplink system using orthogonal propagator method (OPM) over a noisy frequency-selective channel. The performance of typical OPM is limited at low signal-to-noise ratio (SNR) due to its noise-free signal model. The additive white Gaussian noise mainly impacts the diagonal elements of the data covariance matrix (DCM) of the received signal. First, the proposed fractional CFOs estimator minimizes the noise effect by reconstructing the denoised diagonal elements of the DCM for the received signal through cubic Hermite interpolation. Therefore, we use modified OPM to estimate multiple CFOs for MIMO SC-FDMA systems. Second, we have used an iterative method based on the first-order Taylor series approximation of the CFO vector to avoid peak ambiguity and reduce the number of iterations required in the grid search. The mean square error of the proposed CFO estimator considerably improves compared with the existing methods at a low SNR regime. The bit error rate performance of the proposed method for MIMO SC-FDMA is also compared with the existing methods. Finally, the proposed estimator is validated by implementing it on radio frequency hardware testbed over an indoor propagation environment.

Source depth estimation using angle-range interference pattern in deep ocean bottom bounce area

Passive sonar limits the ability to sample vertical scale spatiotemporally, and the received signal is indistinct by noise, resulting in the performance degradation or even failure of the source depth estimation method in deep ocean bottom bounce area. When vertical line array is arranged near the sea surface to locate the source by matched-field processing in bottom bounce area, there is great ambiguity in depth dimension. In this work, the problem of source depth estimation in bottom bounce area is addressed. The peak stripe of angle-range interference pattern is modeled and analyzed based on the ray theory, and a source depth estimation method is proposed, which comprises two parts: estimating sound source depth by utilizing peak stripe fluctuation characteristic combined with spatial spectrum analysis, and reconstructing interference pattern using principal component analysis. The flowchart of the method is listed as follows. Firstly, the spatial spectrum corresponding to each range is spliced to obtain the original angle-range interference pattern. Secondly, the original interference pattern is denoised by principal component analysis to obtain the reconstructed interference structure. Finally, the fluctuation period of peak fringes is extracted from the reconstructed interference pattern to calculate the source depth. Under the low signal-to-noise ratio (SNR) condition, the peak stripe destroyed by noise will reappear in the reconstructed interference pattern and the estimation result of sound source depth will be more accurate, making the source depth estimation method suitable for bottom bounce area. The method combines the space-time cumulative gain of the array and the low-rank characteristic of the interference pattern. The simulation results obtained by Bellhop indicate that the fluctuation period of peak stripe depends on the depth and frequency of the source, which is consistent with the modeling result, and the source depth can be estimated precisely by spatial Fourier transform. The simulation results also show that the rank of the interference pattern is very low and the peak stripe can be composed of a few principal components. Monte Carlo experimental results indicate that the estimated results of the source depth by using reconstructed interference pattern are more accurate than those without using principal component analysis at low SNR condition. The proposed method can achieve more than 80% accuracy at –3 dB SNR.

Seismic Random Noise Attenuation by Applying Multiscale Denoising Convolutional Neural Network

Seismic prospecting is a common method used in oil and gas resource exploration. However, due to the limitations of current collection techniques, seismic records acquired in the field are typically contaminated by severe incoherent noise, which has negative implications for the subsequent processing and interpretation procedures. In addition, numerous traditional denoising algorithms have been applied in order to mitigate this problem, but further improvements are required, especially for the seismic data with spectral overlapping between effective signals and background noise. In recent years, feedforward denoising convolutional neural networks (DnCNNs) have been applied to suppress the complex random noise, and a series of essential insights have been gained. Nonetheless, conventional denoising networks always extract data features depending on single-scale information, resulting in impaired performance when coping with seismic records with a low signal-to-noise ratio (SNR). For solving this problem, a novel multiscale DnCNN (MSDCNN) is developed as an attempt for random noise suppression. Unlike conventional DnCNN, MSDCNN has a hierarchical structure capable of extracting features at different scales and capturing informative and discriminatory features through effective information integration. Meanwhile, the cross-scale feature interaction also increases the processing accuracy when confronted with weak reflection events. Experimental results derived from both synthetic and field data indicate that the proposed network can effectively suppress the random noise and accurately preserve reflection events, even under low SNR conditions.

Improved Moments Estimation for Ground-Based K-Band Doppler Radar Using Cost Function Method

An improved moments estimation algorithm used in a newly developed K-band wind profiler to obtain accurate 3-D profiles of wind velocity along the altitude is described. The signal-to-noise ratio (SNR) of turbulent echoes detected by millimeter-wave radar is usually low, making it becomes a challenging task to determine accurate Doppler profiles. And if wind profilers can provide quick estimates of wind, they will have greater applicability in automated, real-time environments. The improved method combines the concepts of adaptive Doppler windows and profile chain construction, reduces the search range of prospective spectral peaks by using adaptive Doppler windows, and designs new multiparameter cost functions to determine velocity profiles by studying the wind shear in the surface boundary layer. The algorithm can effectively identify the atmospheric echo components of each range bin even under the condition of low SNR, and the computational time is decreased by at least 18% compared with the previous algorithms. The proposed method is tested on the real wind radar datasets to verify its robustness and reliability. The wind information obtained by four different algorithms is compared with the data obtained from the meteorological mast of the Gaoyou wind farm in China. The comparison results indicate that the proposed method shows a nicer match with the data of the mast, and it derives the winds more accurately in the atmosphere of the surface boundary layer.

Super-Resolution Imaging by Computationally Fusing Quantum and Classical Optical Information

A high-speed super-resolution computational imaging technique is introduced on the basis of classical and quantum correlation functions obtained from photon counts collected from quantum emitters illuminated by spatiotemporally structured illumination. The structured illumination is delocalized—allowing the selective excitation of separate groups of emitters as the modulation of the illumination light advances. A recorded set of photon counts contains rich quantum and classical information. By processing photon counts, multiple orders of Glauber correlation functions are extracted. Combinations of the normalized Glauber correlation functions convert photon counts into signals of increasing order that contain increasing spatial frequency information. However, the amount of information above the noise floor drops at higher correlation orders, causing a loss of accessible information in the finer spatial frequency content that is contained in the higher-order signals. We demonstrate an efficient and robust computational imaging algorithm to fuse the spatial frequencies from the low-spatial-frequency range that is available in the classical information with the spatial frequency content in the quantum signals. Because of the overlap of low spatial frequency information, the higher signal-to-noise ratio (SNR) information concentrated in the low spatial frequencies stabilizes the lower SNR at higher spatial frequencies in the higher-order quantum signals. Robust performance of this joint fusion of classical and quantum computational single-pixel imaging is demonstrated with marked increases in spatial frequency content, leading to super-resolution imaging, along with much better mean squared errors in the reconstructed images.

MmWave V2V Localization in MU-MIMO Hybrid Beamforming

Recent trends for vehicular localization in millimetre-wave (mmWave) channels include employing a combination of parameters such as angle of arrival (AOA), angle of departure (AOD), and time of arrival (TOA) of the transmitted/received signals. These parameters are challenging to estimate, which along with the scattering and random nature of mmWave channels, and vehicle mobility lead to errors in localization. To circumvent these challenges, this paper proposes mmWave vehicular localization employing difference of arrival for time and frequency, with multiuser (MU) multiple-input-multiple-output (MIMO) hybrid beamforming; rather than relying on AOD/AOA/TOA estimates. The vehicular localization can exploit the number of vehicles present, as an increase in the number of vehicles reduces the Cramér-Rao bound (CRB) of error estimation. At 10 dB signal-to-noise ratio (SNR) both spatial multiplexing and beamforming result in comparable localization errors. At lower SNR values, spatial multiplexing leads to larger errors compared to beamforming due to formation of spurious peaks in the cross ambiguity function. Accuracy of the estimated parameters is improved by employing an extended Kalman filter leading to a root mean square (RMS) localization error of approximately 6.3 meters.

Speckle-free laser projection structured illumination microscopy based on a digital micromirror device

In the structured illumination microscopy (SIM) family, interferometric implementations using traditional physical gratings or advanced liquid crystal on silicon spatial light modulators are popular but have the disadvantages of expensive and limited speed. A promising alternative of a digital micromirror device (DMD) has improved the situation for its lower cost and higher speed, furthermore, it has contributed to the proposal of a simple and compact projection DMD-SIM. However, the projection DMD-SIM method is accused of erroneously measuring the resolution because of the scattering characteristic of gold beads, despite the inferior experimental resolution that is because of the low fringe contrast and the resulting low signal-to-noise ratio (SNR). As an improvement, we use a high-brightness laser for illumination to achieve higher SNR, and the accompanying problems of speckle and inhomogeneity originating from the laser’s high coherence are solved by a high-frequency phase-randomization deformable mirror in conjunction with a square-core multimode fiber, which does not limit the image acquisition speed and with fewer energy losses compared to the existing methods. Also, in terms of the problem of low SNR, we apply a parallel-iteration Richardson–Lucy deconvolution algorithm for SIM reconstruction to further improve the imaging performance based on ensuring the resolution. Based on these advances, we image fluorescent nanoparticles and other non-biological and biological fluorescent samples, and the experimental results demonstrate the improved imaging performance of our modified system.

The Application of Machine Learning ICA-VMD in an Intelligent Diagnosis System in a Low SNR Environment.

This paper proposes a new method called independent component analysis–variational mode decomposition (ICA-VMD), which combines ICA and VMD. The purpose is to study the application of ICA-VMD in low signal-to-noise ratio (SNR) signal processing and data analysis. ICA is a very important method in the field of machine learning. It is an unsupervised learning algorithm that can dig out the independent factors hidden in the observation signal. The VMD method estimates each signal component by solving the frequency domain variational optimization problem, and it is very suitable for mechanical fault diagnosis. The advantage of ICA-VMD is that it requires two sensory cues to distinguish the original source from the unwanted noise. In the three cases studied here, the original source was first contaminated by white Gaussian noise. The three cases in this study are under different SNR conditions. The SNR in the first case is –6.46 dB, the SNR in the second case is –21.3728, and the SNR in the third case is –46.8177. The simulation results show that the ICA-VMD method can effectively recover the original source from the contaminated data. It is hoped that, in the future, there will be new discoveries and advances in science and technology to solve the noise interference problem through this method.

Spectrogram decomposition of ultrasonic guided waves for cortical thickness assessment using basis learning

Due to its multimode and dispersive nature, ultrasonic guided waves (UGWs) usually consist of overlapped wave packets, which challenge accurate bone characterization. To overcome this obstacle, a classic idea is to separate individual modes and to extract the corresponding dispersion curves. Reported single-channel mode separation algorithms mainly focused on offering a time–frequency representation (TFR) where the energy distributions of individual modes were apart from each other. However, such approaches are still limited to identifying the modes without significant overlapping in time–frequency domain. In this study, a spectrogram decomposition technique was developed based on a combination strategy of generalized separable nonnegative matrix factorization (GS-NMF) and adaptive basis learning, towards the automatic mode extraction under severe overlapping and low signal-to-noise ratio (SNR). The extracted modes were further used for cortical thickness estimation. The method was verified using broadband simulated and experimental datasets. Experiments were conducted on a bone-mimicking plate and bovine cortical bone plates. For simulated data, the relative errors between extracted and theoretical dispersion curves are 1.33% (SNR = ∞), 1.43% (SNR = 10 dB) and 0.88% (SNR = 5 dB). The root-mean-square errors of the estimated thickness for 3.10 mm-thick bone-mimicking plate, 3.83 mm- and 4.00 mm-thick bovine cortical bone plates are 0.039 mm, 0.049 mm, and 0.052 mm, respectively. It is demonstrated that the proposed method is capable of separating multimodal UGWs even under significantly overlapping and low SNR conditions, further facilitating the UGW-based cortical thickness assessment.

DETERMINATION OF FIBER ORIENTATION IN HIGH ANGULAR RESOLUTION DIFFUSION IMAGING USING SPHERICAL HARMONICS AND PARTICLE SWARM OPTIMIZATION

The fiber directions in High Angular Resolution Diffusion Imaging (HARDI) with low fractional anisotropy or low Signal to Noise Ratio (SNR) cannot be estimated accurately. In this paper, the fiber directions are estimated using Particle Swarm Optimization and Spherical Deconvolution (PSO-SD). Fiber orientation is modeled as a Dirac delta function in [Formula: see text]. The Spherical Harmonic Coefficients (SHC) of the Dirac delta function in the [Formula: see text] direction are obtained using the rotational harmonic matrix and the SHC of the Dirac delta function in the [Formula: see text]-axis. The PSO-SD method is used to determine ([Formula: see text]). We generated noise-free synthetic data for isotropic regions (FA varied from 0.1 to 0.8) and synthetic data with two crossing fibers for anisotropic regions with SNRs of 20, 15, 10 and 5 (FA [Formula: see text] 0.78). In the noise-free signal (FA [Formula: see text] 0.3), the Success Ratio (SR) and Mean Difference Angle (MDA) of the PSO-SD method were 1∘ and 9.48∘, respectively. In the noisy signal (FA [Formula: see text] 0.78, SNR [Formula: see text] 10, crossing angle [Formula: see text] 40), the SR and MDA of PSO-SD (with [Formula: see text]) were 0.46∘ and 12.3∘, respectively. The PSO-SD method can estimate fiber directions in HARDI with low fractional anisotropy and low SNR. Moreover, it has a higher SR and lower MDA in comparison with those of the super-CSD method.

Effect of the recording condition on the quality of a single-lead electrocardiogram.

Although many wearable single-lead electrocardiogram (ECG) monitoring devices have been developed, information regarding their ECG quality is limited. This study aimed to evaluate the quality of single-lead ECG in healthy subjects under various conditions (body positions and motions) and in patients with arrhythmias, to estimate requirements for automatic analysis, and to identify a way to improve ECG quality by changing the type and placement of electrodes. A single-lead ECG transmitter was placed on the sternum with a pair of electrodes, and ECG was simultaneously recorded with a conventional Holter ECG in 12 healthy subjects under various conditions and 35 patients with arrhythmias. Subjects with arrhythmias were divided into sinus rhythm (SR) and atrial fibrillation (AF) groups. ECG quality was assessed by calculating the sensitivity and positive predictive value (PPV) of the visual detection of QRS complexes (vQRS), automatic detection of QRS complexes (aQRS), and visual detection of P waves (vP). Accuracy was defined as a 100% sensitivity and PPV. We also measured the amplitude of the baseline, P wave, and QRS complex, and calculated the signal-to-noise ratio (SNR). We then focused on aQRS and estimated thresholds to obtain an accurate aQRS in more than 95% of the data. Finally, we sought to improve ECG quality by changing electrode placement using offset-type electrodes in 10 healthy subjects. The single-lead ECG provided 100% accuracy for vQRS, 87% for aQRS, and 74% for vP in healthy subjects under various conditions. Failure for accurate detection occurred in several motions in which the baseline amplitude was increased or in subjects with low QRS or P amplitude, resulting in low SNR. The single-lead ECG provided 97% accuracy for vQRS, 80% for aQRS in patients with arrhythmias, and 95% accuracy for vP in the SR group. The AF group showed higher baseline amplitude than the SR group (0.08mV vs. 0.02mV, P < 0.01) but no significant difference in accuracy for aQRS (79% vs. 81%, P = 1.00). The thresholds to obtain an accurate aQRS were a QRS amplitude > 0.42mV and a baseline amplitude < 0.20mV. The QRS amplitude was significantly influenced by electrode placement and body position (P < 0.01 for both, two-way analysis of variance), and the maximum reduction by changing body position was estimated as 30% compared to the sitting posture. The QRS amplitude significantly increased when the inter-electrode distance was extended vertically (1.51mV for vertical extension vs. 0.93mV for control, P < 0.01). The single-lead ECG provided at least 97% accuracy for vQRS, 80% for aQRS, and 74% for vP. To obtain stable aQRS in any body positions, a QRS amplitude > 0.60mV and a baseline amplitude < 0.20mV were required in the sitting posture considering the reduction induced by changing body position. Vertical extension of the inter-electrode distance increased the QRS amplitude.

Towards real time guide wire shape extraction in fluoroscopic sequences: A two phase deep learning scheme to extract sparse curvilinear structures.

Real time localization and shape extraction of guide wire in fluoroscopic images plays a significant role in the image guided navigation during cerebral and cardiovascular interventions. Given the complexity of the non-rigid and sparse characteristics of guide wire structures, and the low SNR(Signal Noise Ratio) of fluoroscopic images, traditional handcrafted guide wire tracking methods such as Frangi filter, Hessian Matrix, or open active contour usually produce insufficient accuracy with high computational cost, and may require extra human intervention for proper initialization or correction. The application of deep learning techniques to guide wire tracking is reported to produce significant improvement in guide wire localization accuracy, but the heavy calculation cost is still a concern. In this paper we propose a two phase deep learning scheme for accurate and real time guide wire shape extraction in fluoroscopic sequences. In the first phase we train a guide wire localization network to pick image regions containing guide wire structures. From the picked image regions, we train a guide wire shape extraction network in the second phase to mark the guide wire pixels. We report that our proposed method can accurately distinguish about 99% of the guide wire structure pixels, and the falsely detected pixels in the background are close to 0, the average offset from the ground truth is less than 1 pixel. For extreme cases where traditional handcrafted method may fail, our proposed method can still extract guide wire completely and accurately. The processing time for a 512 × 512 pixels image is 78 ms. Compared with the traditional filtering based method from our previous work, we show that our proposed method can achieve more accurate and stable performance. Compared with other deep learning methods, our proposed method significantly improve calculation efficiency to meet the real time requirement of clinical applications.

Specific emitter identification for IoT devices based on deep residual shrinkage networks

Specific emitter identification can distinguish individual transmitters by analyzing received signals and extracting inherent features of hard-ware circuits. Feature extraction is a key part of traditional machine learning-based methods, but manual extraction is generally limited by prior professional knowledge. At the same time, it has been noted that the performance of most specific emitter identification methods degrades in the low signal-to-noise ratio (SNR) environments. The deep residual shrinkage network (DRSN) is proposed for specific emitter identification, particularly in the low SNRs. The soft threshold can preserve more key features for the improvement of performance, and an identity shortcut can speed up the training process. We collect signals via the receiver to create a dataset in the actual environments. The DRSN is trained to automatically extract features and implement the classification of transmitters. Experimental results show that DRSN obtains the best accuracy under different SNRs and has less running time, which demonstrates the effectiveness of DRSN in identifying specific emitters.

Carbon fiber electrodes for intracellular recording and stimulation

Objective. To understand neural circuit dynamics, it is critical to manipulate and record many individual neurons. Traditional recording methods, such as glass microelectrodes, can only control a small number of neurons. More recently, devices with high electrode density have been developed, but few of them can be used for intracellular recording or stimulation in intact nervous systems. Carbon fiber electrodes (CFEs) are 8 µm-diameter electrodes that can be assembled into dense arrays (pitches ⩾ 80 µm). They have good signal-to-noise ratios (SNRs) and provide stable extracellular recordings both acutely and chronically in neural tissue in vivo (e.g. rat motor cortex). The small fiber size suggests that arrays could be used for intracellular stimulation. Approach. We tested CFEs for intracellular stimulation using the large identified and electrically compact neurons of the marine mollusk Aplysia californica. Neuron cell bodies in Aplysia range from 30 µm to over 250 µm. We compared the efficacy of CFEs to glass microelectrodes by impaling the same neuron’s cell body with both electrodes and connecting them to a DC coupled amplifier. Main results. We observed that intracellular waveforms were essentially identical, but the amplitude and SNR in the CFE were lower than in the glass microelectrode. CFE arrays could record from 3 to 8 neurons simultaneously for many hours, and many of these recordings were intracellular, as shown by simultaneous glass microelectrode recordings. CFEs coated with platinum-iridium could stimulate and had stable impedances over many hours. CFEs not within neurons could record local extracellular activity. Despite the lower SNR, the CFEs could record synaptic potentials. CFEs were less sensitive to mechanical perturbations than glass microelectrodes. Significance. The ability to do stable multi-channel recording while stimulating and recording intracellularly make CFEs a powerful new technology for studying neural circuit dynamics.

Sparse inversion-based seismic random noise attenuation via self-paced learning

Seismic random noise reduction is an important task in seismic data processing at the Chinese loess plateau area, which benefits the geologic structure interpretation and further reservoir prediction. The sparse inversion is one of the widely used tools for seismic random noise reduction, which is often solved via the sparse approximation with a regularization term. The ℓ1 norm and total variation (TV) regularization term are two commonly used regularization terms. However, the ℓ1 norm is only a relaxation of the ℓ0 norm, which cannot always provide a sparse result. The TV regularization term may provide unexpected staircase artifacts. To avoid these disadvantages, we proposed a workflow for seismic random noise reduction by using the self-paced learning (SPL) scheme and a sparse representation (i.e. the continuous wavelet transform, CWT) with a mixed norm regularization, which includes the ℓp norm and the TV regularization. In the implementation, the SPL, which is inspired by human cognitive learning, is introduced to avoid the bad minima of the non-convex cost function. The SPL can first select the high signal-to-noise ratio (SNR) seismic data and then gradually select the low SNR seismic data into the proposed workflow. Moreover, the generalized Beta wavelet (GBW) is adopted as the basic wavelet of the CWT to better match for seismic wavelets and then obtain a more localized time-frequency (TF) representation. It should be noted that the GBW can easily constitute a tight frame, which saves the calculation time. Synthetic and field data examples are adopted to demonstrate the effectiveness of the proposed workflow for effectively suppressing seismic random noises and accurately preserving valid seismic reflections.

A novel method for NMR data denoising based on discrete cosine transform and variable length windows

Raw nuclear magnetic resonance (NMR) data detected exhibits a low signal-to-noise ratio (SNR). The inversion results suffer from the low SNR of echo data. It is therefore essential to denoise the data before NMR data inversion, to improve the accuracy of inverted NMR spectrum. In this paper, a denoising method was firstly proposed based on discrete cosine transform (DCT), whereby four window schemes were designed into one window, multiple non-overlapping windows of equal length, multiple overlapping windows of equal length, and variable length windows. The effectiveness of four schemes was verified and compared by numerical simulations, and it indicates that Denoising Fig. 5 with variable length windows is the optimum. Core experimental data and NMR logging data were used to verify the practicality of Denoising Fig. 5. The results indicated that Denoising Fig. 5 can be effectively used to reduce the noise of actual NMR data, and the processed NMR spectrum with Denoising Fig. 5 is more reliable. • Discrete Cosine Transform was first applied to the noise reduction of NMR echo data. • Variable Length Windows scheme can achieve the best denoising effect. • The proposed method can effectively improve the application effect of NMR logging.

ISAR Imaging Motion Compensation in Low SNR Environments Using Phase Gradient and Filtering Techniques

In a low signal-to-noise ratio (SNR) environment, conventional range profiles-based motion compensation (MoCo) methods become inapplicable for inverse synthetic aperture radar (ISAR) imaging. In this article, a novel method for range profile alignment along with phase adjustment is proposed. The method is applied in the uncompressed phase history domain without the need for a prominent point. Motion errors are estimated from a phase map using only fast Fourier transform (FFT) and Hadamard multiplication operations, making the method suitable for real-time applications. The proposed method strives to aggregate the energy of all scatterers into a single peak by 2-D coherent accumulation incorporating the product of the radar pulse backscattering and its subsequent pulse, hence the reference to phase gradient processing (PGP). This process strengthens SNR and overcomes the problems associated with low signal power, even when the range profiles are submerged in noise. Experimental results are provided to demonstrate the performance of the proposed method compared with the state-of-the-art algorithms.

Real-Time Cerebral Vessel Segmentation in Laser Speckle Contrast Image Based on Unsupervised Domain Adaptation.

Laser speckle contrast imaging (LSCI) is a full-field, high spatiotemporal resolution and low-cost optical technique for measuring blood flow, which has been successfully used for neurovascular imaging. However, due to the low signal–noise ratio and the relatively small sizes, segmenting the cerebral vessels in LSCI has always been a technical challenge. Recently, deep learning has shown its advantages in vascular segmentation. Nonetheless, ground truth by manual labeling is usually required for training the network, which makes it difficult to implement in practice. In this manuscript, we proposed a deep learning-based method for real-time cerebral vessel segmentation of LSCI without ground truth labels, which could be further integrated into intraoperative blood vessel imaging system. Synthetic LSCI images were obtained with a synthesis network from LSCI images and public labeled dataset of Digital Retinal Images for Vessel Extraction, which were then used to train the segmentation network. Using matching strategies to reduce the size discrepancy between retinal images and laser speckle contrast images, we could further significantly improve image synthesis and segmentation performance. In the testing LSCI images of rodent cerebral vessels, the proposed method resulted in a dice similarity coefficient of over 75%.

Experimental evaluation of some current injection-voltage reading patterns in electrical impedance tomography (EIT) and comparison to simulation results - Case study: Large scales

Each electrical impedance tomography system to be prosper, usually would had its own current injection and voltage reading patterns within domains under test. So in different domains, each of sensing patterns have different performances. In industrial application for conductive carrier-phases and in large domains, impedance imaging of real phantoms would be advisable candidate for better performance. Therefore in this study, performance of five sensing patterns in a large scale (outer diameter of 700 mm) were evaluated. Based on the sensing patterns, parameters such as; applicability in the domain and online mode, voltage signal-to-noise ratio (SNR), boundary voltage changes (BVC), sensitivity distributions, size error (SE) and ringing, in different object locations were studied by simulation and experimentally. The results illustrated that there is a significant difference between the reconstruction images of simulation and experimental data in the patterns with low SNR. Also, in a pattern with high SNR, due to the low BVC, the desired performance was not achieved. The highest correlation in the image reconstruction of the simulation and experimental data was obtained in the Opposite-Adjacent pattern with high SNR and BVC. The results show that the selection of this sensing pattern can improve the object's SE up to 0.68%.