Pseudo Points Research Articles

Global phase accuracy enhancement of structured light system calibration and 3D reconstruction by overcoming inevitable unsatisfactory intensity modulation

Mainstream fringe projection profilometry (FPP) faces challenges in measuring low signal-to-noise ratio intensity modulation areas, especially in shadow areas on objects and black areas on calibration plates, causing pseudo phases and obvious nonlinear calibration errors in 3D measurement. To address these issues concurrently, we proposed a cyclic phase-shifts difference (CPSD) module based on phase-shifting gratings. CPSD uses double-layer circulation difference to retain the pixels with higher light intensity eliminating the influence of ambient light and finally creates a mask removing and correcting black and shadow areas. Without further post-processing and projection operation, we achieved the global phase accuracy enhancement and obtained high-precision 3D point clouds without pseudo points. The method’s efficacy was proven through experiments on three kinds of calibration planes, achieving a significant reduction in mean absolute error (MAE) and standard deviation (STD) in plane reconstruction (up to 50.9% and 56.1% respectively). We believe that CPSD has enormous potential for wide application across various FPP-related fields. (Code and data are avaliable at: https://github.com/LiYiMingM/CPSD_shadow_and_black_area_phase_correction_in_SL3D)

PTA-Det: Point Transformer Associating Point Cloud and Image for 3D Object Detection.

In autonomous driving, 3D object detection based on multi-modal data has become an indispensable perceptual approach when facing complex environments around the vehicle. During multi-modal detection, LiDAR and a camera are simultaneously applied for capturing and modeling. However, due to the intrinsic discrepancies between the LiDAR point and camera image, the fusion of the data for object detection encounters a series of problems, which results in most multi-modal detection methods performing worse than LiDAR-only methods. In this investigation, we propose a method named PTA-Det to improve the performance of multi-modal detection. Accompanied by PTA-Det, a Pseudo Point Cloud Generation Network is proposed, which can represent the textural and semantic features of keypoints in the image by pseudo points. Thereafter, through a transformer-based Point Fusion Transition (PFT) module, the features of LiDAR points and pseudo points from an image can be deeply fused under a unified point-based form. The combination of these modules can overcome the main obstacle of cross-modal feature fusion and achieves a complementary and discriminative representation for proposal generation. Extensive experiments on KITTI dataset support the effectiveness of PTA-Det, achieving a mAP (mean average precision) of 77.88% on the car category with relatively few LiDAR input points.

Research on fingerprint image recognition based on convolution neural network

In order to overcome the problem of poor image matching performance of the image recognition method, a method of fingerprint image recognition based on convolution neural network is proposed. In this method, the defaced fingerprint image is pre-processed by smoothing, convergence, equalisation, background foreground segmentation and distortion correction, and the feature points of the defaced fingerprint image are extracted by combining the neighbourhood judgment method, and the information pseudo feature points are removed by fusing the feature points, the centre points are extracted from the feature points of the defaced fingerprint image, and the centre block image is identified by convolution neural network, so as to realise the defaced fingerprint image distinguish. The experimental results show that the performance of restoration and reconstruction is improved. The rejection rate (FRR) is 3.75%, the false recognition rate (FAR) is 1.25%, and the correct recognition rate (CR) is 99.25%.

Clustering Data Streams Based on Shared Density between Micro-Clusters

As more and more applications produce streaming data, clustering data streams has become an important technique for data and knowledge engineering. A typical approach is to summarize the data stream in real-time with an online process into a large number of so called micro-clusters. Micro-clusters represent local density estimates by aggregating the information of many data points in a defined area. On demand, a (modified) conventional clustering algorithm is used in a second offline step to recluster the micro-clusters into larger final clusters. For reclustering, the centers of the micro-clusters are used as pseudo points with the density estimates used as their weights. However, information about density in the area between micro-clusters is not preserved in the online process and reclustering is based on possibly inaccurate assumptions about the distribution of data within and between micro-clusters (e.g., uniform or Gaussian). This paper describes DBSTREAM, the first micro-cluster-based online clustering component that explicitly captures the density between micro-clusters via a shared density graph. The density information in this graph is then exploited for reclustering based on actual density between adjacent micro-clusters. We discuss the space and time complexity of maintaining the shared density graph. Experiments on a wide range of synthetic and real data sets highlight that using shared density improves clustering quality over other popular data stream clustering methods which require the creation of a larger number of smaller micro-clusters to achieve comparable results.

Adaptive slope filtering for airborne Light Detection and Ranging data in urban areas based on region growing rule

Automatic ground filtering for Light Detection and Ranging data is a critical process for the creation of Digital Terrain Models. Several methods have been developed for filtering Light Detection and Ranging data and some improvements have been made but the problem has not been fully solved for the complicated landscape types and abrupt surfaces. In this paper, the region growing algorithm is introduced into the slope-based algorithm. A probabilistic function is introduced and the index of probability is used to further eliminate the pseudo ground points. A comprehensive test of the performance on nine study sites is conducted and compared with the results of seven other published methods based on the Kappa coefficients and the total errors. Overall, the slope-based and region growing algorithm performs the best in urban areas compared with other filtering methods published in terms of average Kappa value and total errors. The slope-based and region growing algorithm is adaptive to the thresholds of the slope and the filtering window, two commonly required parameters. The complexity of terrain does not significantly influence the performance of the slope-based and region growing algorithm, which allows it to be adopted in different terrains for urban areas.

Signal Detection in MIMO-OFDM Systems Based on SSDE Algorithm

Device-to-device communication enables to improve the application performance of multi-input multi-output (MIMO) technology. In a multi-input multi-output orthogonal frequency division multiplexing (MIMO-OFDM) system, its performance is largely reflected by the signal detection algorithm used in receiver. As a sub-optimal maximum likelihood (ML) detection method, the Selective Spanning with Fast Enumeration algorithm can be successfully applied in MIMO-OFDM systems with high-order modulation. However, its Fast Enumeration scheme calculates constellation points based on fixed formula, which tends to yield pseudo constellation points outside of constellation maps, and consequently cannot work well in low-order modulation. To address this, a selective spanning with direct enumeration (SSDE) algorithm is proposed in this paper. Simulation results proved that the SSDE can achieve similar detection performance at a much lower computational cost in comparison with the ML method. The performance in terms of bit error rate (BER) obtained by SSDE method is also superior to those from the Minimum mean square error and Zero forcing detection algorithms with a huge savings in computational load. By adjusting the parameters used in the SSDE, the tradeoff between BER and computation complexity can be flexibly changed to satisfy specific design requirements in different applications.

Pseudo Central-Catheter Method for Asymmetrical Vaginal Prescription with Multi-Channel Cylindrical Applicators in Image-Guided High Dose-Rate Brachytherapy

Multi-channel cylindrical applicators for high dose-rate (HDR) brachytherapy are utilized mainly for special gynecology cases often involving image-guidance high dose-rate (IG-HDR or IGBT) brachytherapy. In these cases, doses are prescribed to partial vaginal wall/depth where the disease is indicated by an MRI study scanned on the same day. These novel IG-HDR procedures are logistical challenges comparing to typical HDR out-patient treatments due to the complexity of planning for asymmetrical prescriptions as well as the extension of the whole procedure. Indeed the main advantages of HDR over LDR brachytherapy would be significantly weaken unless the IG-HDR procedures become more efficient in time, thus timely treatment planning becomes crucial. Based on our clinical experience, we propose the pseudo central-catheter method to reduce errors, to minimize uncertainty and to ensure efficiency. In this method, the central hollow path of the multi-channel cylindrical applicator is digitized into a pseudo catheter similar to other active catheters, and the asymmetrical prescription points could be generated corresponding to the pseudo dwelling points in the pseudo central-catheter. The pseudo dwelling points are later deleted after generating the prescription point arrays. This method is robust with minimal chance of errors or uncertainties, and demonstrates high efficiency with much less chance of uncertainty, which is significant for IG-HDR brachytherapy procedures utilizing multi-channel applicators.

Bayesian site selection for fast Gaussian process regression

Gaussian Process (GP) regression is a popular method in the field of machine learning and computer experiment designs; however, its ability to handle large data sets is hindered by the computational difficulty in inverting a large covariance matrix. Likelihood approximation methods were developed as a fast GP approximation, thereby reducing the computation cost of GP regression by utilizing a much smaller set of unobserved latent variables called pseudo points. This article reports a further improvement to the likelihood approximation methods by simultaneously deciding both the number and locations of the pseudo points. The proposed approach is a Bayesian site selection method where both the number and locations of the pseudo inputs are parameters in the model, and the Bayesian model is solved using a reversible jump Markov chain Monte Carlo technique. Through a number of simulated and real data sets, it is demonstrated that with appropriate priors chosen, the Bayesian site selection method can produce a good balance between computation time and prediction accuracy: it is fast enough to handle large data sets that a full GP is unable to handle, and it improves, quite often remarkably, the prediction accuracy, compared with the existing likelihood approximations.

The Rewards Associated with Meeting or Beating Management Forecasts

This study explores whether investors use management forecasts as earnings benchmarks to evaluate future performance. Our results show that the market reacts positively to firms meeting or beating point forecasts and the two bounds of range forecasts. The findings are more pronounced among firms with higher uncertainty, and robust for a subsample of firms without analyst coverage. Firms meeting or beating these forecasts also demonstrate better future performance. Furthermore, firms manage earnings upward and revise forecasts downward to help meet or beat their forecasts. There is no evidence that the mid-point or pseudo points within range forecasts are valid earnings benchmarks. Overall, our findings suggest that management forecasts (including point forecasts and the two bounds of range forecasts) are earnings benchmarks that help investors to evaluate firm performance. This improves our understanding of the informational role of voluntary disclosures in market pricing.

An approach to simplifying point features on maps using the multiplicative weighted Voronoi diagram

A number of features, such as settlements and islands, are represented using point symbols on intermediate and micro scale maps. If the maps are reduced to smaller scale, the point features need to be simplified to make the maps legible. Hence, it is necessary to develop algorithms for point feature generalisation. For the reason above, an algorithm based on the multiplicatively weighted Voronoi diagram (MWVD) is proposed in the paper. To ensure statistical, thematic, metric and topological information contained in the original point features can be transmitted correctly after simplification, the algorithm selects corresponding measures (i.e. the number of points, weight, Voronoi neighbour, Voronoi polygon and distribution range) to quantify the four types of information, and integrates the measures in the process of point feature generalisation. First, the algorithm detects the range polygon of the given point features; second, it adds the pseudo points (i.e. the vertices of the range polygon) to the original points to form a new point set and tessellates the new point set to get the MWVD; then it computes the selection probability of each point using the area of each Voronoi polygon, and sorts all points in decreasing order by their selection probability values; after this, it marks those to-be-deleted points as ‘deleted’ according to their selection probability values and their Voronoi neighbouring relations, and determines if they can be physically deleted. Finally, the algorithm is ended by comparing the number of points retained on the map with that computed by the Radical Law. The algorithm is parameter-free, automatic and easy to understand, owing to the use of the MWVD. As the experiments show, it can be used in simplification of point features arranged in clusters, such as settlements, islands and control points on topographic maps at intermediate/ micro scale.

An algorithm for point cluster generalization based on the Voronoi diagram

This paper presents an algorithm for point cluster generalization. Four types of information, i.e. statistical, thematic, topological, and metric information are considered, and measures are selected to describe corresponding types of information quantitatively in the algorithm, i.e. the number of points for statistical information, the importance value for thematic information, the Voronoi neighbors for topological information, and the distribution range and relative local density for metric information. Based on these measures, an algorithm for point cluster generalization is developed. Firstly, point clusters are triangulated and a border polygon of the point clusters is obtained. By the border polygon, some pseudo points are added to the original point clusters to form a new point set and a range polygon that encloses all original points is constructed. Secondly, the Voronoi polygons of the new point set are computed in order to obtain the so-called relative local density of each point. Further, the selection probability of each point is computed using its relative local density and importance value, and then mark those will-be-deleted points as ‘deleted’ according to their selection probabilities and Voronoi neighboring relations. Thirdly, if the number of retained points does not satisfy that computed by the Radical Law, physically delete the points marked as ‘deleted’ forming a new point set, and the second step is repeated; else physically deleted pseudo points and the points marked as ‘deleted’, and the generalized point clusters are achieved. Owing to the use of the Voronoi diagram the algorithm is parameter free and fully automatic. As our experiments show, it can be used in the generalization of point features arranged in clusters such as thematic dot maps and control points on cartographic maps.

Construction of device performance models using adaptive interpolation and sensitivities

The performance of a device can be obtained as a continuous function of design parameters by repeated finite element analysis (FEA) followed by interpolation. Since FEA can provide sensitivities at little extra cost, interpolation schemes should make use of them. The scheme proposed here is adaptive and uses radial basis functions (multiquadrics) and pseudo data points. Results are presented for some artificial test problems and for the interpolation of the reflection coefficient of a partial height metallic post in a rectangular waveguide, found by three-dimensional FEA. In every case, the new scheme gives greater accuracy than the equivalent scheme that does not use sensitivities.

Microchemical Oxygen Flask Sulfur Determination: General Review and Collaborative Study of an Alternative Method Using Thorin-Methylene Blue Indicator

Abstract Recent collaborative studies of microchemical oxygen flask sulfur determinations are summarized and compared statistically, and the 1975 collaborative study is presented. This volumetric technique, using a peroxide solution as absorbent, barium perchlorate as titrant, and a combination of thorin and methylene blue as indicator, was tested by 8 collaborators. Data from 72 determinations were reported. The samples studied were sulfanilamide, benzylisothiourea hydrochloride, cystine, potassium sulfate, and potassium sulfate with a phosphate additive. Overall statistical results for the 3 pure organic compounds gave satisfactory values: average standard deviation 0.10, overall average deviation of the mean from the theoretical value 0.13, and average bias −0.04. The overall average deviation of the mean from the theoretical value for the potassium sulfate compound was high (1.94) in this study, and that of the phosphated potassium sulfate compound was even higher (6.46). Thorin-methylene blue does not require a titration assembly (required by tetrahydroxyquinone) and has greater stability than dimethylsulfonazo III whose indicator lag-time causes troublesome pseudo end points. However, a critical factor, when using thorin-methylene blue, is the masking effect of foreign ions. All reagents must be as pure as possible, since concentrations of chloride, fluoride, nitrate, phosphate, potassium, and sodium interfere. The thorin-methylene blue method for sulfur determination has been adopted as official first action as an alternative oxygen flask microchemical method of analysis.