Adversarial attack strategies for 3D object detection have highlighted the critical importance of addressing security concerns in this domain. However, white-box methods require full access to the victim model in large-scale point cloud applications. To this end, we propose a novel Policy-Driven Black-box Attack (BAT) that is designed to optimize attack locations without necessitating detailed knowledge of the victim models. First, we introduce a density-aware pattern generator that creates scene-adaptive attack clusters. Second, we leverage the deep deterministic policy gradient in deep reinforcement learning to train an attack agent capable of targeting the victim model. Ultimately, the attack agent is iteratively directed towards optimal attack locations through the joint application of critic loss and actor loss. To the best of our knowledge, this represents the first reinforcement learning-based black-box attack applied to practical 3D object detection. Experimental results on the KITTI, nuScenes, and Waymo datasets demonstrate that BAT effectively diminishes the accuracy of notable models. Importantly, BAT significantly enhances the attack success rate (surpassing state-of-the-art both white-box and black-box methods) and increases transferability (by 20 times) through simple deep deterministic policy gradient, thus establishing a new baseline for adversarial attacks in 3D object detection.

When the space remote sensing camera is imaging in orbit, the image data usually contain obvious stripe noise and random noise due to the influence of imaging mode (push-broom mode), imaging environment, and other factors. Previous research has primarily concentrated on postprocessing the image using the substantial computational resources of the computer platform, with little consideration given to the feasibility of real-time processing on the satellite platform. In light of the above, a novel hardware platform stripe noise estimation and random noise removal framework (HSENRF) is proposed for the first time, with the objective of enabling real-time simultaneous destriping and denoising on the satellite platform. It employs the theory of maximum likelihood estimation (MLE) to construct the image processing module (IMM) to remove the stripe noise as well as the random noise through the application of the modified nonlocal mean algorithm (M-NLM) as well as the expectation maximization (EM) algorithm. Multiple IMM modules are arranged in the form of a pipeline to achieve multiple iterations of image data. The framework has been successfully deployed on the Xilinx xc7k410tffg676 FPGA platform, and a large number of simulations and experiments show that it has high computational efficiency and good performance. At a system clock of 50 MHz, it achieves a frame rate of 162.77 frames/s when processing 8-bit$512\times 512$image. In addition, it also has strong robustness and can effectively remove noise components in remote sensing images with different scenes and different spatial resolutions.

Single-cell RNA-sequencing (scRNA-seq) has been widely used to depict gene expression profiles at the single-cell resolution. However, its relatively high dropout rate often results in artificial zero expressions of genes and therefore compromised reliability of results. To overcome such unwanted sparsity of scRNA-seq data, several imputation algorithms have been developed to recover the single-cell expression profiles. Here, we propose a novel approach, GE-Impute, to impute the dropout zeros in scRNA-seq data with graph embedding-based neural network model. GE-Impute learns the neural graph representation for each cell and reconstructs the cell-cell similarity network accordingly, which enables better imputation of dropout zeros based on the more accurately allocated neighbors in the similarity network. Gene expression correlation analysis between true expression data and simulated dropout data suggests significantly better performance of GE-Impute on recovering dropout zeros for both droplet- and plated-based scRNA-seq data. GE-Impute also outperforms other imputation methods in identifying differentially expressed genes and improving the unsupervised clustering on datasets from various scRNA-seq techniques. Moreover, GE-Impute enhances the identification of marker genes, facilitating the cell type assignment of clusters. In trajectory analysis, GE-Impute improves time-course scRNA-seq data analysis and reconstructing differentiation trajectory. The above results together demonstrate that GE-Impute could be a useful method to recover the single-cell expression profiles, thus enabling better biological interpretation of scRNA-seq data. GE-Impute is implemented in Python and is freely available at https://github.com/wxbCaterpillar/GE-Impute.

This paper considers constellation designs for 2×2 systems with a coherent dual-polarized antenna configuration. Such an architecture enables a four dimensional (4D) signal space that exploits polarization, and this paper considers hybrid constellations formed as a combination of amplitude/phase states and polarization states. For coherent detection, phase shift keying (PSK) is combined with polarization shift keying (PolSK) to produce constant-power PSK+PolSK constellations that consistently outperform PSK+PSK. Next, quadrature amplitude modulation (QAM) and amplitude phase shift keying (APSK) are combined with PolSK to construct hybrid QAM+PolSK and APSK+PolSK constellations, and constrained capacity results indicate the viability of coded modulation based on these signal spaces. For non-coherent detection, differential PSK (DPSK) is combined with PolSK to yield DPSK+PolSK signaling that does not require phase alignment at the receiver and outperforms DPSK+DPSK, both with and without the adverse effect of carrier frequency offset and symbol time offset. The methods are verified in over-the-air testing with universal software radio peripherals (USRPs) integrated in a dual-polarized testbed. Error performances of all proposed 4D hybrid constellations in an OFDM framework are analyzed under both quasi-AWGN and multipath fading scenarios.

Conventional image classification models commonly adopt a single feature vector to represent informative contents. However, a single image feature system can hardly extract the entirety of the information contained in images, and traditional encoding methods have a large loss of feature information. Aiming to solve this problem, this paper proposes a feature fusion-based image classification model. This model combines the principal component analysis (PCA) algorithm, processed scale invariant feature transform (P-SIFT) and color naming (CN) features to generate mutually independent image representation factors. At the encoding stage of the scale-invariant feature transform (SIFT) feature, the bag-of-visual-word model (BOVW) is used for feature reconstruction. Simultaneously, in order to introduce the spatial information to our extracted features, the rotation invariant spatial pyramid mapping method is introduced for the P-SIFT and CN feature division and representation. At the stage of feature fusion, we adopt a support vector machine with two kernels (SVM-2K) algorithm, which divides the training process into two stages and finally learns the knowledge from the corresponding kernel matrix for the classification performance improvement. The experiments show that the proposed method can effectively improve the accuracy of image description and the precision of image classification.

This paper considers polarization shift keying (PolSK) modulation in wireless communication systems. The PolSK has previously been analyzed in additive white Gaussian noise (AWGN) for optical fiber and wireless communications. In our work, analytical results additionally focus on Rayleigh and Rician fading channels. It is shown that the dual-polarized signaling enjoys an ergodic capacity advantage over co-polarized signaling for a sufficiently large Rician K factor and signal-to-noise ratio. Specific PolSK constellations are formulated and analyzed in AWGN and Rayleigh fading channels. We examine the effect of the channel's polarization parameters on symbol error rates and achievable information. Also, considered are detection candidates to exploit statistical and instantaneous channel state information at the receiver. Over-the-air experiments were conducted to validate the principles. At higher modulation orders, uncoded PolSK shows an error rate advantage over unit symbol power constellations such as phase shift keying (PSK) and differential PSK with/without adverse carrier frequency offsets. The PolSK with dual-polarized antennas also outperforms the same modulation technique implemented with spatially separated co-polarized antenna elements in a line-of-sight (LoS) scenario. The proposed maximum likelihood PolSK detectors exploiting non-Gaussian statistics of the noise on the Poincaré sphere are validated to be superior to zero forcing and minimum mean square error equalization for a non-LoS scenario.

This paper considers the design of signal constellations for a 2 × 2 MIMO digital communication system with dualpolarized transmit and receive antennas. Specifically, it proposes four-dimensional “hybrid” constellations formed as the direct product of an absolute amplitude/phase state and a polarization state. We begin by imposing a constant symbol power constraint and combining phase shift keying (PSK) with polarization shift keying (PolSK) to produce hybrid constellations referred to as PSK+PolSK. As a motivating example, uncoded 8PSK+8PolSK is shown to enjoy a 1.1 dB gain over more conventional 8PSK+8PSK at a symbol error rate of 10-4in additive white Gaussian noise. When the power constraint is removed to form QAM+PolSK an APSK+PolSK constellations, the resulting performance suffers relative to comparable QAM+QAM and APSK+APSK constellations due to the polarization state's sensitivity to symbol power. Capacity calculations provide insight into the viability of coded modulation based on hybrid signaling. Particular hybrid constellations based on QAM+PolSK and APSK+PolSK were shown to provide a capacity advantage over comparable QAM+QAM constellations over a range of rates of practical value.

Summary There are two problems in global over‐complete dictionary: lack of adaptability to image local structure and low computational efficiency. Based on the study of the adaptive sparse representation reconstruction, this paper obtained a series of corresponding sub‐dictionaries by image block subset, then the optimal sub‐dictionary for each reconstruction image block is adaptively selected, which can be more accurately sparse represented modeled to improve the effect and efficiency of the algorithm. In order to promote the ability of sparse representation model, nonlocal self‐similarity prior item is introduced. Meanwhile, the nonlocal self‐similarity model is improved by using the idea of the bilateral filter, and the space distance restraint between pixels is introduced to better keep the image edge information. Moreover, the nonlocal self‐similar distance measure is improved to reduce the amount of calculation. Experimental results show that the proposed algorithm can effectively suppress noise effects and can maintain the image edge details, at the same time, there are certain advantages at both the peak signal to noise ratio (PSNR) and visual effects.

This paper explores fundamental properties of wireless communication using polarization-based modulation. Channel models for polarization shift keying (PolSK) are developed assuming additive white Gaussian noise (AWGN) and Rayleigh fading; the resulting analysis of a maximum likelihood receiver yields a bound on the symbol error rate (SER) for PolSK that is approximately 1.0 dB tighter than the previously published bound. In addition, insight into the fundamental performance of polarization-based modulation is gained by computing a bound on achievable rate derived from the channel's mutual information assuming equiprobable inputs. Finally, it is shown how statistical and instantaneous channel state information (CSI) can be exploited at the receiver to form asymmetric decision regions and improve SER performance.

This paper investigates fundamental properties of polarization-based modulation for wireless communication - and in particular the application of channel coding techniques to such systems. After developing appropriate channel models, bounds on achievable rates are computed, and the performance of exemplary LDPC codes are simulated; this is done for both additive white Gaussian noise channels as well as channels subject to i.i.d. Rayleigh fading. A novel “on/off” modulation scheme is developed that adaptively changes the information-bearing polarization states based on the singular value decomposition (SVD) of the realized channel; this scheme is shown to significantly outperform fixed-constellation schemes as well as adaptive-constellation schemes employing equal energy signals.

In wireless communications, wideband polarization dependent loss (PDL) has become significant with the development of signal processing in wideband polarization domain. Aiming at deriving PDL statistical characteristic, a polarized time-variant multipath channel transfer function matrix H which considers azimuth power spectra, power delay profile (PDP), polarization power imbalance and polarization correlations is proposed. Closed-form expressions have been derived for various autocorrelation functions and cross-correlation functions for the polarized channel to fully characterize its statistical property. PDL is defined as the ratio between maximum and minimum eigenvalues of HHHat frequency f. However, due to the difficulty to analytically derive PDL's statistical characteristic, i.e., the distribution of PDL, numerical sum-of-sinusoids simulator is realized to accurately emulate the channel correlations derived above. Inspired by Weibull distribution, the probability density function (PDF) of PDL is well curve fitted using raw data from numerical simulator. The function is parameterized by scale factor, shape factor and normalization factor which all relate to polarization power imbalance and polarization correlations of the polarized channel. Results show that the degree of attenuation unbalance towards eigen-polarizations is in the descending order as: NLOS macrocell > NLOS microcell > NLOS picocell.

In this paper, we propose a new blind spectrum sensing method based on the polarization characteristic of the received signal, which is completely represented by the orientation of a polarization vector. We first discuss a spectrum sensing model based on polarization vectors' orientation. Then we develop the directional statistics of polarization vectors that contain both the signal and noise or noise only. The distinctive difference between the two statistics can be used to decide whether the primary signal exists or not. Based on this, by using the well-known generalized likelihood ratio test (GLRT) paradigm, a new polarization sensing algorithm GLRT-polarization vector (GLRT-PV) is proposed. By applying directional statistics, we derive closed-form expressions for the probability of false alarm and the probability of detection under both dual-polarized additive white Gaussian noise (AWGN) and Rayleigh-fading channels. Our numerical simulation and experimental results show that the proposed method exhibits better performance than other existing methods in the case of unknown primary transmitter polarization and/or presence of noise power uncertainty.

Web service composition is a new paradigm to develop distributed and reactive software-intensive systems. Predicting and preventing failures of dynamic composite services is an important and challenge problem due to the dynamically evolving attribute. In previous work, we propose CASSANDRA, a novel proactive monitoring technique with the ability to predict and prevent the potential failures happening in dynamic evolvable system. In this paper, we concretize the approach into web service composition field. By combining runtime information and design-time specification of basic services, the approach can analyse future -step models ahead of the current service execution states. Then, this model can be used to check with a set of desired properties represented by property sequence chart. Initial experiments on an online medicine case study validates our approach and shows encouraging results.

A tight estimate on the maximum distance between a subdivision surface and its linear approximation is introduced to guide adaptive subdivision with guaranteed accuracy.

Abstract Accurate and robust interference detection and ray‐tracing of subdivision surfaces requires safe linear approximations. Approximation of the limit surface by the subdivided control polyhedron can be both inaccurate and, due to the exponential growth of the number of facets, costly. This paper shows how a standard intersection hierarchy, such as an OBB tree, can be made safe and efficient for subdivision surface interference detection. The key is to construct, on the fly, optimally placed facets, whose spherical offsets tightly enclose the limit surface. The spherically offset facets can be locally subdivided and they can be efficiently intersected based on standard triangle‐triangle interference detection. Categories and Subject Descriptors (according to ACM CCS): I.3.5 [Computer Graphics]: Computational Geometry and Object Modeling