Channel Maps Research Articles

Self-Matching CAM: A Novel Accurate Visual Explanation of CNNs for SAR Image Interpretation

Synthetic aperture radar (SAR) image interpretation has long been an important but challenging task in SAR imaging processing. Generally, SAR image interpretation comprises complex procedures including filtering, feature extraction, image segmentation, and target recognition, which greatly reduce the efficiency of data processing. In an era of deep learning, numerous automatic target recognition methods have been proposed based on convolutional neural networks (CNNs) due to their strong capabilities for data abstraction and mining. In contrast to general methods, CNNs own an end-to-end structure where complex data preprocessing is not needed, thus the efficiency can be improved dramatically once a CNN is well trained. However, the recognition mechanism of a CNN is unclear, which hinders its application in many scenarios. In this paper, Self-Matching class activation mapping (CAM) is proposed to visualize what a CNN learns from SAR images to make a decision. Self-Matching CAM assigns a pixel-wise weight matrix to feature maps of different channels by matching them with the input SAR image. By using Self-Matching CAM, the detailed information of the target can be well preserved in an accurate visual explanation heatmap of a CNN for SAR image interpretation. Numerous experiments on a benchmark dataset (MSTAR) verify the validity of Self-Matching CAM.

Observational Signature of Tightly Wound Spirals Driven by Buoyancy Resonances in Protoplanetary Disks

Abstract Besides the spirals induced by the Lindblad resonances, planets can generate a family of tightly wound spirals through buoyancy resonances. The excitation of buoyancy resonances depends on the thermal relaxation timescale of the gas. By computing timescales of various processes associated with thermal relaxation, namely, radiation, diffusion, and gas–dust collision, we show that the thermal relaxation in protoplanetary disks’ surface layers (Z/R ≳ 0.1) and outer disks (R ≳ 100 au) is limited by infrequent gas–dust collisions. The use of the isothermal equation of state or rapid cooling, common in protoplanetary disk simulations, is therefore not justified. Using three-dimensional hydrodynamic simulations, we show that the collision-limited slow thermal relaxation provides favorable conditions for buoyancy resonances to develop. Buoyancy resonances produce predominantly vertical motions, whose magnitude at the 12CO emission surface is of the order of 100 m s−1 for Jovian-mass planets, sufficiently large to detect using molecular line observations with ALMA. We generate synthetic observations and describe characteristic features of buoyancy resonances in Keplerian-subtracted moment maps and velocity channel maps. Based on the morphology and magnitude of the perturbation, we propose that the tightly wound spirals observed in TW Hya could be driven by a (sub-)Jovian-mass planet at 90 au. We discuss how non-Keplerian motions driven by buoyancy resonances can be distinguished from those driven by other origins. We argue that observations of multiple lines tracing different heights, with sufficiently high spatial/spectral resolution and sensitivity to separate the emission arising from the near and far sides of the disk, will help constrain the origin of non-Keplerian motions.

The theory of kinks – I. A semi-analytic model of velocity perturbations due to planet–disc interaction

ABSTRACT A new technique to detect protoplanets is by observing the kinematics of the surrounding gas. Gravitational perturbations from a planet produce peculiar ‘kinks’ in channel maps of different gas species. In this paper, we show that such kinks can be reproduced using semi-analytic models for the velocity perturbation induced by a planet. In doing so we (i) confirm that the observed kinks are consistent with the planet-induced wake; (ii) show how to quantify the planet mass from the kink amplitude; in particular, we show that the kink amplitude scales with the square root of the planet mass for channels far from the planet velocity, steepening to linear as the channels approach the planet; and (iii) show how to extend the theory to include the effect of damping, which may be needed in order to have localized kinks.

Vehicle Localization via Cooperative Channel Mapping

This paper addresses vehicle positioning, a topic whose importance has risen dramatically in the context of future autonomous driving systems. While classical methods that use GPS and/or beacon signals from network infrastructure for triangulation tend to be sensitive to multi-paths and signal obstruction, our method exhibits robustness with respect to such phenomena. Our approach builds on the recently proposed Channel-SLAM method which first enabled leveraging of multi-path so as to improve (single) vehicle positioning. Here, we propose a cooperative mapping approach which builds upon the Channel-SLAM concept, referred to here as Team Channel-SLAM. Team Channel-SLAM not only exploits the stationary nature of many reflecting objects around the vehicle, but also capitalizes on the multi-vehicle nature of road traffic. The key intuition behind our method is the exploitation for the first time of the correlation between reflectors around multiple neighboring vehicles. An algorithm is derived for reflector selection and estimation, combined with a team particle filter (TPF) so as to achieve high precision simultaneous multiple vehicle positioning. We obtain large improvement over the single-vehicle positioning scenario, with gains being already noticeable for moderate vehicle densities, such as over <b xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">40</b> <b xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</b> improvement for a vehicle density as low as 4 vehicles in 132 meters' length road.

Single Versus Multi-channel Dispersion Analysis of Ultrasonic Guided Waves Propagating in Long Bones.

Ultrasonic guided wave techniques have been applied to characterize cortical bone for osteoporosis assessment. Compared with the current gold-standard X-ray-based diagnostic methods, ultrasound-based techniques pose some advantages such as compactness, low cost, lack of ionizing radiation, and their ability to detect the mechanical properties of the cortex. Axial transmission technique with a source-receiver offset is employed to acquire the ultrasound data. The dispersion characteristics of the guided waves in bones are normally analyzed in the transformed domains using the dispersion curves. The transformed domain can be time-frequency map using a single channel or wavenumber-frequency (or phase velocity-frequency) map with multi-channels. In terms of acquisition effort, the first method is more cost- and time-effective than the latter. However, it remains unclear whether single-channel dispersion analysis can provide as much quantitative guided-wave information as the multi-channel analysis. The objective of this study is to compare the two methods using numerically simulated and ex vivo data of a simple bovine bone plate and explore their advantages and disadvantages. Both single- and multi-channel signal processing approaches are implemented using sparsity-constrained optimization algorithms to reinforce the focusing power. While the single-channel data acquisition and processing are much faster than those of the multi-channel, modal identification and analysis of the multi-channel data are straightforward and more convincing.

Technique for Separating Velocity and Density Contributions in Spectroscopic Data and Its Application to Studying Turbulence and Magnetic Fields

Abstract Based on the theoretical description of position–position–velocity (PPV) statistics in Lazarian &amp; Pogosyan, we introduce a new technique called the velocity decomposition algorithm (VDA) for separating the PPV fluctuations arising from velocity and density fluctuations. Using MHD turbulence simulations, we demonstrate its promise in retrieving the velocity fluctuations from the PPV cube in various physical conditions and its prospects in accurately tracing the magnetic field. We find that for localized clouds, the velocity fluctuations are most prominent in the wing part of the spectral line, and they dominate the density fluctuations. The same velocity dominance applies to extended H i regions undergoing galactic rotation. Our numerical experiment demonstrates that velocity channels arising from the cold phase of atomic hydrogen (H i) are still affected by velocity fluctuations at small scales. We apply the VDA to H i GALFA-DR2 data corresponding to the high-velocity cloud HVC186+19-114 and high-latitude galactic diffuse H i data. Our study confirms the crucial role of velocity fluctuations in explaining why linear structures are observed within PPV cubes. We discuss the implications of VDA for both magnetic field studies and predicting polarized galactic emission that acts as the foreground for cosmic microwave background studies. Additionally, we address the controversy related to the filamentary nature of the H i channel maps and explain the importance of velocity fluctuations in the formation of structures in PPV data cubes. VDA will allow astronomers to obtain velocity fluctuations from almost every piece of spectroscopic PPV data and allow direct investigations of the turbulent velocity field in observations.

A highly non-Keplerian protoplanetary disc

Context. In recent years high-angular-resolution observations have revealed that circumstellar discs appear in a variety of shapes with diverse substructures being ubiquitous. This has given rise to the question of whether these substructures are triggered by planet–disc interactions. Besides direct imaging, one of the most promising methods to distinguish between different disc-shaping mechanisms is to study the kinematics of the gas disc. In particular, the deviations of the rotation profile from Keplerian velocity can be used to probe perturbations in the gas pressure profile that may be caused by embedded (proto-) planets. Aims. In this paper we aim to analyse the gas brightness temperature and kinematics of the transitional disc around the intermediate-mass star CQ Tau in order to resolve and characterise substructure in the gas caused by possible perturbers. Methods. For our analysis we used spatially resolved ALMA observations of the three CO isotopologues 12CO, 13CO, and C18O (J = 2−1) from the disc around CQ Tau. We further extracted robust line centroids for each channel map and fitted a number of Keplerian disc models to the velocity field. Results. The gas kinematics of the CQ Tau disc present non-Keplerian features, showing bent and twisted iso-velocity curves in 12CO and 13CO. Significant spiral structures are detected between ~10 and 180 au in both the brightness temperature and the rotation velocity of 12CO after subtraction of an azimuthally symmetric model, which may be tracing planet–disc interactions with an embedded planet or low-mass companion. We identify three spirals, two in the brightness temperature and one in the velocity residuals, spanning a large azimuth and radial extent. The brightness temperature spirals are morphologically connected to spirals observed in near-infrared scattered light in the same disc, indicating a common origin. Together with the observed large dust and gas cavity, these spiral structures support the hypothesis of a massive embedded companion in the CQ Tau disc.

Uncorrelated feature encoding for faster image style transfer

Recent image style transfer methods use a pre-trained convolutional neural network as their feature encoder. However, the pre-trained network is not optimal for image style transfer but rather for image classification. Furthermore, they require time-consuming feature alignment to consider the existing correlation among channels of the encoded feature map. In this paper, we propose an end-to-end learning method that optimizes both encoder and decoder networks for style transfer task and relieves the computational complexity of the existing correlation-aware feature alignment. First, we performed end-to-end learning that updates not only decoder but also encoder parameters for the task of image style transfer in the network training phase. Second, in addition to the previous style and content losses, we use uncorrelation loss, i.e., the total correlation coefficient among responses of encoder channels. Our uncorrelation loss allows the encoder network to generate a feature map of channels without correlation. Subsequently, our method results in faster forward processing with only a light-weighted transformer of correlation-unaware feature alignment. Moreover, our method drastically reduced the channel redundancy of the encoded feature during the network training process. This provides us a possibility to perform channel elimination with negligible degradation in generated style quality. Our method is applicable to multiple scaled style transfer by using the cascade network scheme and allows a user to control style strength through the usage of a content-style trade-off parameter.

LCANet: Learnable Connected Attention Network for Human Identification Using Dental Images.

Forensic odontology is regarded as an important branch of forensics dealing with human identification based on dental identification. This paper proposes a novel method that uses deep convolution neural networks to assist in human identification by automatically and accurately matching 2-D panoramic dental X-ray images. Designed as a top-down architecture, the network incorporates an improved channel attention module and a learnable connected module to better extract features for matching. By integrating associated features among all channel maps, the channel attention module can selectively emphasize interdependent channel information, which contributes to more precise recognition results. The learnable connected module not only connects different layers in a feed-forward fashion but also searches the optimal connections for each connected layer, resulting in automatically and adaptively learning the connections among layers. Extensive experiments demonstrate that our method can achieve new state-of-the-art performance in human identification using dental images. Specifically, the method is tested on a dataset including 1,168 dental panoramic images of 503 different subjects, and its dental image recognition accuracy for human identification reaches 87.21% rank-1 accuracy and 95.34% rank-5 accuracy. Code has been released on Github. (https://github.com/cclaiyc/TIdentify).

Mud stream formation conditions and mud stream hazardous areas of the Chechen Republic

Mud streams in the mountainous zone are formed almost everywhere in the Chechen Republic: on the northern slope of the Main Ridge and its spurs, the Skalisty, Pastbischny, Lesisty, Terskiy, Sunzhensky ranges and in the depressions between them. The conditions for the mud streams formation and mud stream hazardous regions of the Chechen Republic are determined by orographic tectonic-geomorphological, geological-lithological, soil-plant, anthropogenic and especially hydrometeorological conditions and affect the mud streams with their indicators. The variety and complexity of the relief, as well as the temperature regime, are used to predict the mud stream threat and develop anti-mud stream measures. The authors of the article provide the following indicators: air temperature regime, also values of average monthly and annual air temperatures, monthly values of vertical air temperature gradients in the Chechen Republic territory, changes in annual precipitation (mm) with terrain altitude above the sea level on the northern slope of the Eastern Caucasus, schematic map of orohydrography and mud stream channels of the Chechen Republic on a scale -1: 500,000.

Att-Net: Enhanced emotion recognition system using lightweight self-attention module

Speech emotion recognition (SER) is an active research field of digital signal processing and plays a crucial role in numerous applications of Human–computer interaction (HCI). Nowadays, the baseline state of the art systems has quite a low accuracy and high computations, which needs upgrading to make it reasonable for real-time industrial uses such as detection of content from speech data. The main intent for low recognition rate and high computational cost is a scarceness of datasets, model configuration, and patterns recognition that is the supreme stimulating work for building a robust SER system. In this study, we address these problems and propose a simple and lightweight deep learning-based self-attention module (SAM) for SER system. The transitional features map is given to SAM, which produces efficiently the channel and spatial axes attention map with insignificant overheads. We use a multi-layer perceptron (MLP) in channel attention to extracting global cues and a special dilated convolutional neural network (CNN) in spatial attention to extract spatial info from input tensor. Moreover, we merge, spatial and channel attention maps to produce a combine attention weights as a self-attention module. We placed SAM in the middle of convolutional and connected layers and trained it in an end-to-end mode. The ablation study and comprehensive experimentations are accompanied over IEMOCAP, RAVDESS, and EMO-DB speech emotion datasets. The proposed SER system shows consistent improvements in overall experiments for all datasets and shows 78.01%, 80.00%, and 93.00% average recall, respectively.

The CARMA-NRO Orion Survey: Filament Formation via Collision-induced Magnetic Reconnection—the Stick in Orion A

Abstract A unique filament is identified in the Herschel maps of the Orion A giant molecular cloud. The filament, which we name the Stick, is ruler-straight and at an early evolutionary stage. Transverse position–velocity diagrams show two velocity components closing in on the Stick. The filament shows consecutive rings/forks in C18O (1−0) channel maps, which is reminiscent of structures generated by magnetic reconnection. We propose that the Stick formed via collision-induced magnetic reconnection (CMR). We use the magnetohydrodynamics code Athena++ to simulate the collision between two diffuse molecular clumps, each carrying an antiparallel magnetic field. The clump collision produces a narrow, straight, dense filament with a factor of &gt;200 increase in density. The production of the dense gas is seven times faster than freefall collapse. The dense filament shows ring/fork-like structures in radiative transfer maps. Cores in the filament are confined by surface magnetic pressure. CMR can be an important dense-gas-producing mechanism in the Galaxy and beyond.

An Effective Dual Self-Attention Residual Network for Seizure Prediction.

As one of the most challenging data analysis tasks in chronic brain diseases, epileptic seizure prediction has attracted extensive attention from many researchers. Seizure prediction, can greatly improve patients' quality of life in many ways, such as preventing accidents and reducing harm that may occur during epileptic seizures. This work aims to develop a general method for predicting seizures in specific patients through exploring the time-frequency correlation of features obtained from multi-channel EEG signals. We convert the original EEG signals into spectrograms that represent time-frequency characteristics by applying short-time Fourier transform (STFT) to the EEG signals. For the first time, we propose a dual self-attention residual network (RDANet) that combines a spectrum attention module integrating local features with global features, with a channel attention module mining the interdependence between channel mappings to achieve better forecasting performance. Our proposed approach achieved a sensitivity of 89.33%, a specificity of 93.02%, an AUC of 91.26% and an accuracy of 92.07% on 13 patients from the public CHB-MIT scalp EEG dataset. Our experiments show that different EEG signal prediction segment lengths are an important factor affecting prediction performance. Our proposed method is competitive and achieves good robustness without patient-specific engineering.

Testing Disk Identification Methods through Numerical Simulations of Protostellar Evolution

Abstract We test whether the radii of circumstellar disks can be reliably determined in observations by applying the results of a numerical simulation. First, we execute a core-collapse simulation, which starts from a rotating magnetized spherical core, and continue the calculation until the protostellar mass reaches 0.5 M ⊙. Then, for each set of simulation data, we calculate the radiative transfer to generate the data cube for the synthetic observation. The spatial and velocity resolutions of the synthetic observation are 0.″15 (20 au) and 0.1 km s−1, respectively. We define seven different disk radii. Four radii are estimated from the synthetic observation, using the continuum image, continuum visibility, C18O channel map, and C18O position–velocity (PV) diagram. The other three radii are taken from the simulation and use the disk rotation, infall motion, and density contrast around the protostar to identify the disk. Finally, we compare the disk radii estimated from the systemic observation with those from the simulation. We find that the disk radius defined using the PV diagram can reliably trace the Keplerian disk when the protostellar mass is larger than M * ≳ 0.2 M ⊙, independent of the inclination angle to the line of sight. In addition, the PV diagram provides an accurate estimate of the central stellar mass through the whole protostellar evolution. The simulation also indicates that the circumstellar disk is massive enough to be gravitationally unstable through the evolution. Such an unstable disk can show either a circular or spiral morphology on a similar timescale.

An Efficient Cooperative Routing Algorithm for Large-Scale Wireless Networks based on Capacity and Location

Cooperative communication has been extensively used in small-scale wireless networks, by virtue of its capability to mitigate the fading effects. Extending cooperative communication technique to large-scale networks undergoes severe interference; multi-radio-multi-channel is exploited in large scale wireless networks to ameliorate throughput by concurrent transmissions over orthogonal channels to dwindle interference. Capitalising on these potential advantages, stand in need for efficacious routing and potent mapping of channels to radios. In this paper, we have proffered an algorithm by concurrently considering: 1) relay selection using distance and maximum-available-channel-capacity parameters; 2) channel assignment using dynamic global table. Our algorithm selects a relay-node based on the available channel capacity in the shortest-path to minimise the transmission delay without affecting aggregate throughput. The evaluation and analysis of the proposed model is compared with interference-aware-cooperative-routing algorithm and observed that the proposed algorithm has improved delay reduction by 27.5% and the average number of hops by 32.5%.

Untangling Magnetic Complexity in Protoplanetary Disks with the Zeeman Effect

Abstract With the recent advent of circular polarization capabilities at the Atacama Large Millimeter/submillimeter Array (ALMA), Zeeman effect measurements of spectral lines are now possible as a means to directly probe line-of-sight magnetic fields in protoplanetary disks (PPDs). We present a modeling study that aims to guide physical interpretation of these anticipated observations. Using a fiducial density structure based on a typical ringed disk, we simulate line emission for the hyperfine components of the CN J = 1−0 transition with the POLARIS radiative transfer code. Since the expected magnetic field and typical CN distribution in PPDs remain largely unconstrained, we produce models with several different configurations. Corresponding integrated Stokes I and V profiles and 0.4 km s−1 resolution, 1″ beam convolved channel maps are presented. We demonstrate that the emission signatures from toroidally dominated magnetic fields are distinguishable from vertically dominated magnetic field based on channel map morphology. Due to line-of-sight and beam cancellation effects, disks with toroidal -field configurations result in significantly diminished Stokes V emission. Complex magnetic fields therefore render the traditionally used method for inferring line-of-sight magnetic field strengths (i.e., fitting the derivative of the Stokes I to the Stokes V profile) ambiguous, since a given intrinsic field strength can yield a variety of Stokes V amplitudes depending on the magnetic field geometry. In addition, gas gaps can create structure in the integrated Stokes V profile that might mimic magnetic substructure. This method should therefore be applied with caution in PPD environments and can only confidently be used as a measure of magnetic field strength if the disk’s magnetic field configuration is well understood.

Similarity-Based Prediction for Channel Mapping and User Positioning

In a wireless network, gathering information at the base station about mobile users based only on uplink channel measurements is an interesting challenge. Indeed, accessing the users locations and predicting their downlink channels would be particularly useful in order to optimize the network efficiency. In this letter, a supervised machine learning approach addressing these tasks in an unified way is proposed. It relies on a labeled database that can be acquired in a simple way by the base station while operating. The proposed regression method can be seen as a computationally efficient two layers neural network initialized with a non-parametric estimator. It is illustrated on realistic channel data, both for the positioning and channel mapping tasks, achieving better results than previously proposed approaches, at a lower cost.

Gradient of tactile properties in the rat whisker pad.

The array of vibrissae on a rat's face is the first stage in a high-resolution tactile sensing system. Progressing from rostral to caudal in any vibrissae row results in an increase in whisker length and thickness. This may, in turn, provide a systematic map of separate tactile channels governed by the mechanical properties of the whiskers. To examine whether this map is expressed in a location-dependent transformation of tactile signals into whisker vibrations and neuronal responses, we monitored whiskers' movements across various surfaces and edges. We found a robust rostral-caudal (R-C) gradient of tactile information transmission in which rostral shorter vibrissae displayed a higher sensitivity and bigger differences in response to different textures, whereas longer caudal vibrissae were less sensitive. This gradient is evident in several dynamic properties of vibrissae trajectories. As rodents sample the environment with multiple vibrissae, we found that combining tactile signals from multiple vibrissae resulted in an increased sensitivity and bigger differences in response to the different textures. Nonetheless, we found that texture identity is not represented spatially across the whisker pad. Based on the responses of first-order sensory neurons, we found that they adhere to the tactile information conveyed by the vibrissae. That is, neurons innervating rostral vibrissae were better suited for texture discrimination, whereas neurons innervating caudal vibrissae were more suited for edge detection. These results suggest that the whisker array in rodents forms a sensory structure in which different facets of tactile information are transmitted through location-dependent gradient of vibrissae on the rat's face.

Performance and optimization of hybrid FSO/RF communication system in varying weather

Radio frequency (RF) spectrum is already dense enough and hard to add more broadband channels to meet the current user demands. Optical free-space communications could be an excellent alternative to the RF communications system, and it provides additional benefits, e.g., large bandwidth, high data rates and reliable communication link. Therefore, free-space optical (FSO) communication system becomes more attractive for the deployment of additional broadband channels, and it fulfils the current user demands of bandwidth-hungry applications. FSO communication links are susceptible to numerous meteorological conditions such as fog, snow, dust, smoke, scintillation and smog. Achieving better connectivity under the above-mentioned severe conditions is a crucial research question. Joint optical-RF communication system is developed to overcome the problems as mentioned earlier. The proposed adaptive optical-RF transmission system is optimized so that the system performance is maximized under all channel conditions. Optimization is achieved over the respective channel mappings, and the total required power by exploiting the proposed algorithm. The mapping schemes of each link are optimally chosen such that the total mutual information is maximized while distributing optimal power to the individual channel. Simulations are performed and verified with the analytical results to validate the proposed design. A comparison of the adaptive joint system (i.e., hybrid FSO-RF) over the non-adaptive system under various weather conditions is provided. From simulation results, the performance gain of more than 1 dB is achieved under the minimum power level. It is, therefore, recommended that the adaptive hybrid FSO-RF communication system is always an optimum solution for all weather conditions.

Multi-temporal UAV based repeat monitoring of rivers sensitive to flood

ABSTRACT Multi-temporal repeat monitoring of flood-vulnerable rivers is crucial due to rapid alteration of morphological properties of in-channel landforms. Besides, the characteristics of the river crossing bridges may deteriorate due to flood induced scouring around bridge piles or due to flood loads. Thus, in this study high-resolution topography of the study region was acquired during two consecutive years by unmanned aerial vehicle (UAV) based surveys using Structure-from-Motion (SfM) processing. Following the extraction of digital elevation models (DEM), repeat data that were obtained at each UAV survey were compared using Geomorphic Change Detection (GCD) to calculate volumes of deposition and erosion via DEM of difference (DoD) algorithm. Thus, detailed high-resolution maps of the river channels can be rapidly and efficiently generated by low cost UAV based measurement methods in order for continuous tracking of stream channel morphology for the rivers sensitive to floods.