Low Overlap Ratios Research Articles

Dose-efficient automatic differentiation for ptychographic reconstruction

Ptychography, as a powerful lensless imaging method, has become a popular member of the coherent diffractive imaging family over decades of development. The ability to utilize low-dose X-rays and/or fast scans offers a big advantage in a ptychographic measurement (for example, when measuring radiation-sensitive samples), but results in low-photon statistics, making the subsequent phase retrieval challenging. Here, we demonstrate a dose-efficient automatic differentiation framework for ptychographic reconstruction (DAP) at low-photon statistics and low overlap ratio. As no reciprocal space constraint is required in this DAP framework, the framework, based on various forward models, shows superior performance under these conditions. It effectively suppresses potential artifacts in the reconstructed images, especially for the inherent periodic artifact in a raster scan. We validate the effectiveness and robustness of this method using both simulated and measured datasets.

D3Former: Jointly learning repeatable dense detectors and feature-enhanced descriptors via saliency-guided transformer

Establishing accurate and representative matches is a crucial step in addressing the point cloud registration problem. A commonly employed approach involves detecting keypoints with salient geometric features and subsequently mapping these keypoints from one frame of the point cloud to another. However, methods within this category are hampered by the repeatability of the sampled keypoints. In this paper, we introduce a saliency-guided transformer, referred to as D3Former, which entails the joint learning of repeatable Dense Detectors and feature-enhanced Descriptors. The model comprises a Feature Enhancement Descriptor Learning (FEDL) module and a Repetitive Keypoints Detector Learning (RKDL) module. The FEDL module utilizes a region attention mechanism to enhance feature distinctiveness, while the RKDL module focuses on detecting repeatable keypoints to enhance matching capabilities. Extensive experimental results on challenging indoor and outdoor benchmarks demonstrate that our proposed method consistently outperforms state-of-the-art point cloud matching methods. Notably, tests on 3DLoMatch, even with a low overlap ratio, show that our method consistently outperforms recently published approaches such as RoReg and RoITr. For instance, with the number of extracted keypoints reduced to 250, the registration recall scores for RoReg, RoITr, and our method are 64.3%, 73.6%, and 76.5%, respectively.

Numerical investigation into performance of Savonius wind turbine equipped with inner blades: Overlap ratio effect

This research focuses on the effect of overlap ratio on the performance of Savonius rotor in the presence of inner blades in such a way that inner blade tip is parallel to the main blade tip and root. Findings reveal that at low overlap ratios, using inner blade that the inner blade tip is parallel to the main blade root leads to a higher power and torque coefficients than the other inner blade configuration and also conventional rotor. This research indicates that for the overlapped rotor and for the two inner blade configurations, the higher power and torque coefficients could be obtained at high and low tip speed ratios, respectively. Results show that for both inner blade configurations, the overlapped rotors overcome the negative torque generated by the non-overlapped one. It is also shown that at high overlap ratio, using inner blade has no advantage over the conventional rotor.

Unsupervised Pixel-Level Detection of Rail Surface Defects Using Multistep Domain Adaptation

Detection of rail surface defects using deep learning largely relies on supervision with pixel-level ground truths, where the detection accuracy suffers due to unseen image domains. In order to streamline the laborious labeling process and accurately identify defects on rail surfaces a U-Net-based network that integrates spatial and channel attention (SA) modules, multistep domain adaptation (MSDA), and progressive histogram matching (PHM) is developed in the context of pixel-level defects detection in this research. MSDA with an unlabeled intermediate dataset is used to reduce the domain gap underlying different domains. To further improve the adaptation of the transferred model, the PHM method and the self-training method are employed to enhance detection accuracy. Trained in the source domain with ground truths, the deep learning model is then tested in the target domain with no annotations. Experimental results on the RSDD dataset and the Rail-Joint dataset demonstrate the applicability and effectiveness of the proposed approach. By alleviating the domain discrepancy, the developed model drastically enhances the average precision (AP) score of unsupervised detection from 0.103 to 0.861, which shows competitive performance compared with the AP score of 0.895 by the state-of-the-art supervised models. The novelty of this research lies in the development of a hybrid deep learning approach that is able to achieve competitive detection accuracy with a low overlap ratio between domains and few data labeling.

Dynamic regional splitting planning of remote sensing satellite swarm using parallel genetic PSO algorithm

The strips from region splitting method are critical premise of meta-task generation and task planning for satellite swarm. To achieve the regional observation with high coverage and low overlap ratio and satellites usage, a novel region splitting algorithm is designed based on Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) algorithm. In the hybrid algorithm of GA and PSO, these two algorithms are executed paralleled, and then exchange the best genes and individuals periodically, to overcome the disadvantage of GA's poor local searching ability and PSO's low convergence speed. The encoding method and parallel iterative optimal method are given in detail. Finally, a simulation scenario is designed and executed to verify proposed algorithm, and the effects of algorithm parameters are analyzed by simulation results.

Eliminating poverty through development: The dynamic evolution of multidimensional poverty in rural China

This paper measures intertemporal multidimensional poverty in rural China from 1988 to 2018 based on the China Household Income Project (CHIP) data and for that time period has the following findings. First, multidimensional poverty in rural China declines steadily when measured through the income poverty approach. Second, multidimensional poverty in rural areas presents demographic and regional heterogeneity, among which elderly poverty, child poverty, and poverty in the western Chinese region are the most prominent ones. Third, income poverty and multidimensional poverty have a low overlap ratio and show intertemporal synchronisation over time. Although increases in income can alleviate multidimensional poverty, this effect is very limited. This paper suggests that a suitable multidimensional poverty identification framework should be developed in the era of poverty alleviation after 2020 and that in the new poverty alleviation governance system of China, targeted development strategies should be formulated with an increased focus on the poverty of special groups such as children and the elderly.

Unsupervised SAR and Optical Image Matching Using Siamese Domain Adaptation

Due to its highly complementary information about remote sensing, synthetic aperture radar (SAR) and optical imagery matching have drawn much attention in recent years. Compared with traditional methods, deep learning-based SAR-optical image matching models largely rely on supervision with ground truths, where the matching accuracy suffers because of unseen image domains. To mitigate loads in burdensome labeling tasks, transferring deep learning models trained with annotated source domains to nonannotated target domains has attracted great concern. Due to the domain gap, the difference between the source and target domains is likely to deteriorate the matching accuracy on target data if the training process is directly conducted without proper domain adaptation (DA). In this research, a Siamese DA (SDA) approach with a combined loss function is developed in the context of multimodality image matching. Then, a novel rotation/scale-invariant transformation module with regression modules is designed to extract rotation/scale-equivariant features. Finally, the causal inference-based self-learning method and the multiresolution histogram matching approach are employed to enhance the unsupervised matching performance. Experimental results on the RadarSat/Planet dataset and the Sentinel-1/2 dataset demonstrate that the developed model can achieve competitive matching performance with a low overlap ratio between domains and little data labeling. By alleviating the domain discrepancy, the developed model drastically reduces the average L2 score of the unsupervised matching from 9.576 to 0.658, while the less-than-one-pixel matching error rate is enhanced from 66.3% to 90.6%.

Experimental research on increasing the static torque for a small Savonius rotor of helical type

A small Savonius wind turbine of helical type was tested in an open jet wind facility for identifying the optimum position of the rotor blades aiming the maximization of static torque. The two blades of the wind turbine of U shape were twisted up to 1800 from bottom to top and three values for blade overlap were considered i.e. 0.35, 0.2 and 0.0. The turbine torque was measured for no load conditions and wind velocities between 2-8 m/s. In the same time the air flow through Savonius rotor was observed by using a visualization setup aiming to identify the correlation between the torque and the flow pattern inside the returning blade at rotating angles where the torque has highest values. Best results for the torque coefficient of 0.72 and power coefficient of 0.265 were obtained at a wind velocity of 5.8 m/s and low overlap ratio of 0.2. By measuring the overall torque of the rotor for each case study, the positive effect of pressure recovery was explained as a result of air flow through the blade overlap from the concave side of the advancing blade to concave side of the returning blade for a certain interval of the rotating angle a = 50-60°.

An Accelerated and Robust Partial Registration Algorithm for Point Clouds

Partial registration for point clouds plays an important role in various fields such as 3D mapping reconstruction, remote sensing, unmanned driving, and cultural heritage protection. Unfortunately, partial registration is challenging due to difficulties such as the low overlap ratio of two point clouds and the perturbation in the orderless and sparse 3D point clouds. Thus, a variety of the 3D shape context descriptors are introduced for finding the optimal matching. However, extracting geometric features and descriptors are time consuming and easily degenerated by noise. To overcome these problems, we introduce a parallel coarse-to-fine partial registration method. Our contributions can be summarized as: Firstly, a robust coarse trimmed method is proposed to estimate the coarse overlap area and the initial transformation via fast bilateral denoising and parallel point feature histogram (PPFH) descriptor aligning. Secondly, an accelerated fine registration procedure is conducted by a parallel trimmed iterative closest point (PTrICP) method. Moreover, most parts of our coarse-to-fine workflow are accelerated under the Graphics Processing Unit (GPU) parallel execution mode for efficiency. Thirdly, we extend our method from the rigid registration to the isotropic scaling registration, which improves its applicability. Experiments have demonstrated that our method is feasible and robust in various situations, including the low overlap ratio, outlier, noise and scaling.

Comparison of progesterone concentrations in blubber and plasma among female Antarctic minke whales of known reproductive status

The utility of progesterone concentration in blubber as a means of determining reproductive status in the Antarctic minke whale Balaenoptera bonaerensis was assessed through a comparative analysis of progesterone concentration in blubber and plasma among 230 female whales of known reproductive status (immature, resting, ovulating or pregnant). Whales were sampled during the austral summer in the Antarctic Ocean. The general pattern of progesterone concentration by reproductive category was well correlated between blubber and plasma samples, validating in principle the use of progesterone concentrations in blubber to determine the reproductive status of females. However, some differences were found for resting and ovulating females, which require further consideration. For blubber, overlap of progesterone concentrations was observed between reproductive categories with the exception of immature/ovulating and immature/pregnant. This result suggests that the method of using progesterone concentration in blubber cannot distinguish between pregnant and non-pregnant mature females. However, it can be used to distinguish between immature and mature females. Although a low overlap ratio in concentration was found between immature and resting females, the method is still useful for determining sexual maturity, because resting females of the Antarctic minke whale are seldom found in the Antarctic Ocean.

A Modified Method for Registration of 3D Point Clouds with a Low Overlap Ratio

본 논문에서는 노이즈를 포함한 채 오버랩 영역이 적은 두 점군을 정합할 때 정확도와 수렴 속도를 향상시키는 알고리즘을 제시한다. 정확도를 높이기 위하여 점군의 기하학 정보를 최대한 활용하며, 정합 단계에서는 노이즈가 포함된 점군에서 오버랩 되는 영역을 적절히 선택하고, 개선된 가속 알고리즘을 사용하여 정합 속도를 향상시킨다. 정확도를 향상시키는 기존의 방법은 노이즈가 많은 점군에 적용할 수 없으므로, 본 논문에서는 정합에 사용되는 영역을 선택하는 것으로써 기존 방법의 문제를 해결하였다. 또한 똑같은 점군쌍에서만 적용되는 가속 알고리즘을 낮은 오버랩의 점군쌍에 적용하였다. 기존의 방법에 간단한 알고리즘을 추가함으로써 서너 배 더 빠른 수렴 속도를 낼 수 있도록 하였다. 결론적으로, 노이즈가 많고 오버랩이 적은 점군쌍의 정합에 있어서 본 논문에서 제시하는 알고리즘을 적용하면 속도와 정확도가 향상되었음을 알 수 있다.

Nonrigid Image Registration for Low-Altitude SUAV Images With Large Viewpoint Changes

Low-altitude aerial photography using small unmanned aerial vehicles (SUAVs) with large viewpoint changes causes nonrigid distortions and low overlap ratios. We present a nonrigid feature-based low-altitude SUAV image-registration method. The key idea of our method is to maintain a high matching ratio on inliers while taking advantage of outliers for varying the warping grids. Thus, accurate image transformation over the overlapping areas as well as a good approximation of the real transformation over the nonoverlapping areas can be obtained. Experiments on feature matching and image registration are performed using 42 pairs of SUAV images. Our method exhibited a favorable performance as compared with four state-of-the-art methods, even with up to 80% outliers.

Deep Learning-based MSMS Spectra Reduction in Support of Running Multiple Protein Search Engines on Cloud.

The diversity of the available protein search engines with respect to the utilized matching algorithms, the low overlap ratios among their results and the disparity of their coverage encourage the community of proteomics to utilize ensemble solutions of different search engines. The advancing in cloud computing technology and the availability of distributed processing clusters can also provide support to this task. However, data transferring and results' combining, in this case, could be the major bottleneck. The flood of billions of observed mass spectra, hundreds of Gigabytes or potentially Terabytes of data, could easily cause the congestions, increase the risk of failure, poor performance, add more computations' cost, and waste available resources. Therefore, in this study, we propose a deep learning model in order to mitigate the traffic over cloud network and, thus reduce the cost of cloud computing. The model, which depends on the top 50 intensities and their m/z values of each spectrum, removes any spectrum which is predicted not to pass the majority voting of the participated search engines. Our results using three search engines namely: pFind, Comet and X!Tandem, and four different datasets are promising and promote the investment in deep learning to solve such type of Big data problems.

Ptychographic Algorithm Using Dual-Tree Complex Wavelet Transform

Reconstructing the interesting complex image from multiple diffraction patterns is the goal of the ptychography. Previous ptychographic algorithms often suffer from low reconstruction quality under the low overlap ratios. To address this issue, we proposed a novel ptychographic phase retrieval (PR) algorithm of exploiting the sparsity of the image in dual-tree complex wavelet domain. The Fourier magnitude measurements are utilized to construct a data fidelity term, and the sparse representation model of the image over the dual-tree complex wavelet transform is utilized for the sparse induced regularization term. The data fidelity term and the proposed regularization term are combined to formulate a ptychographic PR optimization problem. Alternating direction method of multipliers (ADMM) and gradient descent algorithm are utilized for solving the corresponding optimization problem. Compared with previous algorithms, the experimental results indicate that the proposed algorithm can obtain reconstructed images with high quality even at low overlap ratios.

The plant virus microscope image registration method based on mismatches removing

The electron microscopy is one of the major means to observe the virus. The view of virus microscope images is limited by making specimen and the size of the camera's view field. To solve this problem, the virus sample is produced into multi-slice for information fusion and image registration techniques are applied to obtain large field and whole sections. Image registration techniques have been developed in the past decades for increasing the camera's field of view. Nevertheless, these approaches typically work in batch mode and rely on motorized microscopes. Alternatively, the methods are conceived just to provide visually pleasant registration for high overlap ratio image sequence. This work presents a method for virus microscope image registration acquired with detailed visual information and subpixel accuracy, even when overlap ratio of image sequence is 10% or less. The method proposed focus on the correspondence set and interimage transformation. A mismatch removal strategy is proposed by the spatial consistency and the components of keypoint to enrich the correspondence set. And the translation model parameter as well as tonal inhomogeneities is corrected by the hierarchical estimation and model select. In the experiments performed, we tested different registration approaches and virus images, confirming that the translation model is not always stationary, despite the fact that the images of the sample come from the same sequence. The mismatch removal strategy makes building registration of virus microscope images at subpixel accuracy easier and optional parameters for building registration according to the hierarchical estimation and model select strategies make the proposed method high precision and reliable for low overlap ratio image sequence.

Interference suppression for the LTE uplink

Interference cancellation is expected to have significant importance for next-generation wireless communication systems due to various co-channel deployment scenarios and denser frequency reuse. In this study, an interference cancellation receiver that exploits the unique characteristics of single-carrier frequency-division multiple access based systems is proposed. The proposed receiver suppresses the co-channel dominant interference by blanking frequency-domain samples where the desired and interfering signals overlap. In order to improve the performance, demodulation and regeneration stages can be introduced and repeated multiple times. Further enhancement is possible by initially accommodating a group of reliable symbols before the iterations. The simulation results indicate that proposed methods work particularly well for low overlap ratios compared to interference coordination and no cancellation schemes.

Third body effects in the wear of polyamide: Micro-mechanisms and wear particles analysis

The wear micro-mechanisms at a frictional interface between a polyamide 4,6 (PA 46) substrate and a spherical sapphire counterface have been investigated using in situ contact observation. Experiments were carried out under reciprocating sliding conditions with varying values of the overlap ratio, Δ, defined as the ratio of the displacement amplitude to the contact diameter. Using this procedure, it was found possible to vary the extent of wear particles accumulation within coherent third body agglomerates. For low overlap ratios ( Δ < 0.2), particles detached from the PA 46 substrate were extensively agglomerated within distinct third body corrugations. In this regime, the rate limiting process regarding the wear degradation was found to be the extrusion of fibril-like wear particles from the extremities of sheared third body agglomerates. On the other hand, no substantial debris accumulation was found to occur when the overlap ratio was greater than 0.2. Additional infrared thermal contact imaging showed that these wear mechanisms were induced under essentially isothermal conditions. From differential scanning calorimetry (DSC) and size exclusion chromatography (SEC) analysis, micro-structural changes within wear particles were found to be associated with a strong reduction in the molecular weight of the PA 46 molecules by virtue of stress-induced chain scission mechanisms.

Linear Clustering Algorithm by Competitive Learning Suitable for Coal Mine in Wireless Sensor Networks

Because of the complex environment of coal mines, there are many accidents like communication link fault caused by mobile equipment and sliding stones. Therefore, fast rebuilding of remaining nodes and dynamic adding of nodes toted by the rescuer are required after the disaster. Furthermore, the sensor nodes used for localization of the miner and his equipment can only be deployed redundantly along the straight laneway to form a linear wireless sensor network, which proposes a new requirement of topology control strategy. Combining with the self-organizing learning theory of the artificial neural networks, LCACL is presented, a linear clustering algorithm by competitive learning, in which sensor nodes are divided into mobile nodes which add or leave the network dynamically and fixed nodes which could act as sleep nodes, followers, cluster-bridges, or cluster-heads self-adaptively. On the comprehensive consideration of the remaining energy of the sensor node and the number of neighboring nodes, the competition vector, the base for competition, and evaluation vector, the evaluation standard, are formed, which correspond to the input vector and weight vector of artificial neural networks respectively. A similarity comparison using the Euclidean distance is made between them after normalization, and the competitive selection of roles and conversion are achieved using the Euclidean distance. Only the winner node will be entitled to make a learning adjustment, and then broadcasts the adjusted competition vector, which is the evaluation vector in the next competition. According to the rule of cluster-head selection from odd hops and cluster-bridge selection from even hops, the network is initialized as that nodes are divided into clusters of scale-uniform and low overlap ratio. The cluster-head initiates the competitive selection of the sleep node if there are too many followers, and its followers compete to sleep. Along with the energy consumption of the node, the cluster-head or cluster-bridge initiates a new election if it detects its energy consumption exceeding a certain threshold, and all the nodes in the cluster participate in competition by means of competitive learning to achieve role conversion. The simulation results show that the network organized by LCACL has balanced energy consumption, stronger ability to resist destruction, better expansibility, and longer lifetime of network, and the ratio of overlap could be 14% if the communication radius is 70 meters.