S^(2)ANet:Combining local spectral and spatial point grouping for point cloud processing

Background Despite the recent progress in 3D point cloud processing using deep convolutional neural networks,the inability to extract local features remains a challenging problem.In addition,existing methods consider only the spatial domain in the feature extraction process.Methods In this paper,we propose a spectral and spatial aggregation convolutional network(S^(2)ANet),which combines spectral and spatial features for point cloud processing.First,we calculate the local frequency of the point cloud in the spectral domain.Then,we use the local frequency to group points and provide a spectral aggregation convolution module to extract the features of the points grouped by the local frequency.We simultaneously extract the local features in the spatial domain to supplement the final features.Results S^(2)ANet was applied in several point cloud analysis tasks;it achieved stateof-the-art classification accuracies of 93.8%,88.0%,and 83.1%on the ModelNet40,ShapeNetCore,and ScanObjectNN datasets,respectively.For indoor scene segmentation,training and testing were performed on the S3DIS dataset,and the mean intersection over union was 62.4%.Conclusions The proposed S^(2)ANet can effectively capture the local geometric information of point clouds,thereby improving accuracy on various tasks.

Continuous time-varying Q-factor estimation method in the time-frequency domain

The Q-factor is an important physical parameter for characterizing the absorption and attenuation of seismic waves propagating in underground media,which is of great signifi cance for improving the resolution of seismic data,oil and gas detection,and reservoir description.In this paper,the local centroid frequency is defi ned using shaping regularization and used to estimate the Q values of the formation.We propose a continuous time-varying Q-estimation method in the time-frequency domain according to the local centroid frequency,namely,the local centroid frequency shift(LCFS)method.This method can reasonably reduce the calculation error caused by the low accuracy of the time picking of the target formation in the traditional methods.The theoretical and real seismic data processing results show that the time-varying Q values can be accurately estimated using the LCFS method.Compared with the traditional Q-estimation methods,this method does not need to extract the top and bottom interfaces of the target formation;it can also obtain relatively reasonable Q values when there is no eff ective frequency spectrum information.Simultaneously,a reasonable inverse Q fi ltering result can be obtained using the continuous time-varying Q values.

An Efficient Topography Adaptive Filter for Insar Processing

An efficient implementation of the topography adaptive filter based on local frequency estimation is proposed, where chirp z transform is applied to enhance the accuracy of the frequency estimation. As a by product of this adaptive filter, the linear approximated phase model of the interferogram is employed to improve the coherence estimation. The impacts of the adaptive filter on global and local phase unwrapping algorithms are discussed. Finally, aiming at the negative effect that the adaptive filter can bring to local phase unwrapping algorithms, a fusion scheme that takes advantage of least square and several local phase unwrapping algorithms is presented.

Mechanism of Pendulum-type wave phenomenon in deep block rock mass

Pendulum-type ( μ wave) wave is a new type of elastic wave propagated with low frequency and low velocity in deep block rock masses. The μ wave is sharply different from the traditional longitudinal and transverse waves propagated in continuum media and is also a phenomenon of the sign-variable reaction of deep block rock masses to dynamic actions, besides the Anomalous Low Friction (ALF) phenomenon. In order to confirm the existence of the μ wave and study the rule of variation of this μ wave experimentally and theoretically, we first carried out one-dimensional low-speed impact experiments on granite and cement mortar blocks and continuum block models with different characteristic dimensions, based on the multipurpose testing system developed by us independently. The effects of model material and dimensions of models on the propagation properties of 1D stress wave in blocks medium are discussed. Based on a comparison and analysis of the propagation properties (acceleration amplitudes and Fourier spectra) of stress wave in these models, we conclude that the fractures in rock mass have considerable effect on the attenuation of the stress wave and retardarce of high frequency waves. We compared our model test data with the data of in-situ measurements from deep mines in Russia and their conclusions. The low-frequency waves occurring in blocks models were validated as Pendulum-type wave. The frequencies corresponding to local maxima of spectral density curves of three-directional acceleration satisfied several canonical sequences with the multiple of 2～(1/2), most of those frequencies satisfied the quantitative expression (2～(1/2))i V p/2△ .

Directional Spectrum of Wind Waves: Part Ⅱ- Comparison and Confirmation

A model on the directional frequency spectrum of wind waves for deep water is introduced. The comparisons of the proposed model with other existing models show that the proposed model is very close to the JONSWAP model and DHH model for describing the developing waves under the normal spectral bandwidth, and has a better description for the transition of the unidirectional spectrum from ω -4 to ω -5 at a position around 3ω p, i.e., three time the peak frequency. Comparisons also show that the proposed model describes closely both field data measured by a four-frequency radar and a laser-optical sensor, and laboratory data measured by a laser slope gauge and an imaging optical method. The comparisons further demonstrate that the inverse spectral bandwidth as a new wave parameter is robust for describing the spectral steepness. Finally, the formula on the local spectral-peak angular frequency is confirmed using the observed two-dimensional spectra.

Directional Spectrum of Wind Waves: Part Ⅰ- Model and Derivation

A new model on the directional spectrum of wind waves for deep water is proposed based on the statistics of wind waves. This model contains three parameters: the wave age, the inverse spectral bandwidth and the local spectral-peak angular frequency. The inverse spectral bandwidth is a robust parameter for describing the spectral steepness of wind waves. Using the inverse spectral bandwidth parameter, the proposed model can well describe various observations obtained from the open ocean and laboratory tank.

High-Speed Numeral Filter in Impulse Voltage Measurement

Since the CPU of embed system has some limitation in operating speed, a new filter was put forward which implemented mountain template convolution by performing rectangle template convolution two times. It can obtain time and frequency localization with computational complexity greatly reduced. This algorithm was applied to lightning waveforms (include chopped waveforms) parameter calculation. It simplifies the computation and the results pretreated by this algorithm are in accord with IEC1083-2 completely. It was applied in embed system successfully. Its capability in frequency restraining was researched. The validity of the algorithm was proved in theory when processing lightning waves. The standard sources and the processing results are consistent completely.

Local Region Frequency Guided Dynamic Inconsistency Network for Deepfake Video Detection

In recent years,with the rapid development of deepfake technology,a large number of deepfake videos have emerged on the Internet,which poses a huge threat to national politics,social stability,and personal privacy.Although many existing deepfake detection methods exhibit excellent performance for known manipulations,their detection capabilities are not strong when faced with unknown manipulations.Therefore,in order to obtain better generalization ability,this paper analyzes global and local inter-frame dynamic inconsistencies from the perspective of spatial and frequency domains,and proposes a Local region Frequency Guided Dynamic Inconsistency Network(LFGDIN).The network includes two parts:Global SpatioTemporal Network(GSTN)and Local Region Frequency Guided Module(LRFGM).The GSTN is responsible for capturing the dynamic information of the entire face,while the LRFGM focuses on extracting the frequency dynamic information of the eyes and mouth.The LRFGM guides the GTSN to concentrate on dynamic inconsistency in some significant local regions through local region alignment,so as to improve the model's detection performance.Experiments on the three public datasets(FF++,DFDC,and Celeb-DF)show that compared with many recent advanced methods,the proposed method achieves better detection results when detecting deepfake videos of unknown manipulation types.

Approximate Maximum Likelihood Algorithm for Moving Source Localization Using TDOA and FDOA Measurements

A closed-form approximate maximum likelihood（AML） algorithm for estimating the position and velocity of a moving source is proposed by utilizing the time difference of arrival（TDOA） and frequency difference of arrival（FDOA） measurements of a signal received at a number of receivers.The maximum likelihood（ML） technique is a powerful tool to solve this problem.But a direct approach that uses the ML estimator to solve the localization problem is exhaustive search in the solution space,and it is very computationally expensive,and prohibits real-time processing.On the basis of ML function,a closed-form approximate solution to the ML equations can be obtained,which can allow real-time implementation as well as global convergence.Simulation results show that the proposed estimator achieves better performance than the two-step weighted least squares（WLS） approach,which makes it possible to attain the Cramér-Rao lower bound（CRLB） at a sufficiently high noise level before the threshold effect occurs.

Partial Regularity for the 3D Magneto-hydrodynamics System with Hyper-dissipation

We prove that for the 3D MHD equations with hyper-dissipations （-△）^α （ 1 〈 α 〈 5/4） the Hausdorff dimension of singular set at the first blowing up time is at most 5 - 4α, by means of physical and frequency localization, Bony＇s paraproduct and Littlewood-Paley theory.