Two-stage Method Research Articles

The Effects Of Early And Delayed Palatal Repair On Maxillary Growth And Speech Outcome In Unilateral Cleft Lip And Palate Patients: A Systematic Review

Some cleft clinics use a two-stage method to correct the cleft palate, which helps avoid reduced maximum growth after early hard palate surgery and impaired speech development after late soft palate repair.. This study sought to evaluate the effects of early vs late palatoplasty on speech and maximum growth results. The following databases were searched: Cochrane Library, Science Direct, Pubmed, and NCBI. The primary objective of this study was to determine whether unilateral, non-syndromic cleft lip, alveolus, and palate patients who underwent early or late hard palate surgery before or after 18 months of age had an impact on speech and maxillary development. Things like case studies, reviews, editorials, textbooks, and studies conducted on animals weren't included. The following systematic observations consist of eight studies. Most studies had methodological quality that ranged from fair to good. This review discusses the experience of cleft centers using a variety of one and two stage palatoplasty procedures and timing, and shows inconsistent results regarding maximal progression. The seemingly inconsistent results of the studies included in this observation of speech outcomes suggest that there is insufficient evidence to conclude whether one- or two-stage palatoplasty has a significant effect on maximal development for a variety of reasons. The review provides inconclusive evidence on the impact of hard palate repair timing on facial growth and speech outcomes in unilateral cleft lip and palate patients, highlighting the need for further controlled studies.

Sources of specialist physician fee variation: Evidence from Australian health insurance claims data

This study explores the variation in specialist physician fees and examines whether the variation can be attributed to patient risk factors, variation between physicians, medical specialties, or other factors. We use health insurance claims data from a large private health insurer in Australia. Although Australia has a publicly funded health system that provides universal health coverage, about 44 % of the population holds private health insurance. Specialist physician fees in the private sector are unregulated; physicians can charge any price they want, subject to market forces.We examine the variation in fees using two price measures: total fees charged and out-of- pocket payments. We follow a two-stage method of removing the influence of patient risk factors by computing risk-adjusted prices at patient-level, and aggregating the adjusted prices over all claims made by each physician to arrive at physician-level average prices. In the second stage, we use variance-component models to analyse the variation in the physician-level average prices.We find that patient risk factors account for a small portion of the variance in fees and out-of-pocket payments. Physician-specific variation accounts for the bulk of the vari- ance. The results underscore the importance of understanding physician characteristics in formulating policy efforts to reduce fee variation.

Large-Scale Land Cover Mapping Framework Based on Prior Product Label Generation: A Case Study of Cambodia

Large-Scale land cover mapping (LLCM) based on deep learning models necessitates a substantial number of high-precision sample datasets. However, the limited availability of such datasets poses challenges in regularly updating land cover products. A commonly referenced method involves utilizing prior products (PPs) as labels to achieve up-to-date land cover mapping. Nonetheless, the accuracy of PPs at the regional level remains uncertain, and the Remote Sensing Image (RSI) corresponding to the product is not publicly accessible. Consequently, the sample dataset constructed through geographic location matching may lack precision. Errors in such datasets are not only due to inherent product discrepancies, and can also arise from temporal and scale disparities between the RSI and PPs. In order to solve the above problems, this paper proposes an LLCM framework for generating labels for use with PPs. The framework consists of three main parts. First, initial generation of labels, in which the collected PPs are integrated based on D-S evidence theory and initial labels are obtained using the generated trust map. Second, for dynamic label correction, a two-stage training method based on initial labels is adopted. The correction model is pretrained in the first stage, then the confidence probability (CP) correction module of the dynamic threshold value and NDVI correction module are introduced in the second stage. The initial labels are iteratively corrected while the model is trained using the joint correction loss, with the corrected labels obtained after training. Finally, the classification model is trained using the corrected labels. Using the proposed land cover mapping framework, this study used PPs to produce a 10 m spatial resolution land cover map of Cambodia in 2020. The overall accuracy of the land cover map was 91.68% and the Kappa value was 0.8808. Based on these results, the proposed mapping framework can effectively use PPs to update medium-resolution large-scale land cover datasets, and provides a powerful solution for label acquisition in LLCM projects.

Intelligent design and optimization system for shear wall structures based on large language models and generative artificial intelligence

Intelligent design technology for shear wall structures has great potential for enhancing design efficiency and addressing the challenges of tedious and repetitive design tasks. Recently, there has been a surge in the development of this technology. However, existing deep learning-based design methods for shear wall structures suffer from poor quality and usability issues. To address these challenges, this study proposes an intelligent design and optimization system for shear wall structures based on large language models (LLMs) and generative artificial intelligence (AI). The system employs an LLM as the core controller, which interacts with engineers to interpret their language descriptions and translate them into executable computer code. Subsequently, the system utilizes the corresponding structural generation and optimization methods to accomplish intelligent design tasks automatically. Furthermore, the system incorporates such critical factors as the empirical rules, mechanical performance, and material consumption into the structural optimization process. A unique three-level, two-stage optimization method is constructed based on topology, pattern, and size to enhance the overall design quality. Being able to complete the entire workflow of architectural drawing processing, structural scheme generation, and analysis model establishment, the proposed system enables automated and efficient design of shear wall structures. Through the analysis and validation of multiple cases, it was demonstrated that this system can significantly speed up the design by approximately 30 times compared to that of traditional methods whilst ensuring the safety and cost-effectiveness of the design schemes. Consequently, this study provides valuable insights for the advancement of automated structural design undertakings.

Landslide Damming Threats Along the Jinsha River, China

The 2280 km long Jinsha River has been blocked at least four times in the past 30 years. A landslide damming hazard chain can endanger communities and infrastructures hundreds of kilometers downstream from the damming site in alpine gorges. Past damming events have resulted in severe consequences, demanding a thorough assessment of damming threats along the entire Jinsha River. This study digitizes the Jinsha River and visualizes its topographic, tectonic, hydrologic, and climate characteristics in detail. A two-stage full-probability method is proposed for assessing the damming threats along this river, making it possible to identify potential damming hotspots and high-priority zones for hazard mitigation. It is found that the upper reach of the Jinsha River poses the greatest damming threat, and the threat level gradually decreases downstream. Approximately 33.4%, 36.7%, 20.5%, and 9.4% of the entire length of the Jinsha River are classified as low, moderate, high, and very high threat levels, respectively. Compared with existing hydropower projects, future projects in the upper reach are more likely to be exposed to landslide damming. We highlight the value of basin-scale spatial threat analysis and envisage that our findings will promote more targeted local-scale risk assessments for potential damming hotspots. These outcomes provide the basis for managing the risks of river damming and hydropower infrastructure along the Jinsha River.

Robust Assessment of Hosting Capacity of Renewable Generation in Distribution Systems Considering the Active Network Management

Abstract As the penetration rate of renewable distributed generation (RDG) increases, its random fluctuation characteristics have brought many problems to the safe operation of the active distribution network (ADN), such as voltage over-limit, reverse flow, and deterioration of power quality. Therefore, it has effectively affected the new energy hosting capacity of the ADN. Therefore, the effective assessment of the new energy hosting capacity of the ADN is a key part of the distribution network planning. In order to cope with the random fluctuation characteristics of new energy, this paper presents a two-stage robust assessment method of RDG hosting capacity of distribution networks by considering active network management (ANM) techniques such as network reconfiguration, reactive power compensation and on-load voltage regulation. Finally, the improved IEEE-33 distribution system is used to verify the results. The proposed method has been shown to be effective for evaluating the ADN’s RDG hosting capacity.

Method for Two-Stage Radar Measurement of the Blade Repetition Rate of a Propeller-Driven Aircraft

Introduction. Radar images of aircraft propellers can significantly improve the quality of their recognition and protection against simulating interference. Such images can be obtained using algorithms based on an inverse antenna aperture synthesis. A key factor determining the quality of image acquisition is the accuracy of the rotor blade rotation frequency measurement. In 2019, a method for measuring the blade repetition rate was proposed, which is based on convolution of the "secondary" signal modulation spectrum while simultaneously eliminating the influence of the Doppler frequency of the signal reflected from the aircraft body. In sequential analysis, the number of cycles is determined by the ratio of the maximum blade repetition rate (hundreds of hertz) to the discrete frequency shift (thousandths of hertz). In this case, to solve the measurement problem, the required number of cycles should be hundreds of thousands, which is expensive in terms of practical implementation.Aim. Development of a two-stage method for measuring the blade repetition rate, which allows the number of signal convolution cycles to be reduced by hundreds of times.Materials and methods. The proposed method is aimed at implementing adaptation circuits to an a priori unknown rotor speed of an aircraft, which can be determined based on the blade rotation frequency. The method involves measuring the blade frequency in two stages: a rough measurement of the blade frequency rate followed by its accurate measurement within the limits of the maximum errors of the rough measurement.Results. A method for a two-stage measurement of the blade repetition rate as applied to the construction of radar images of aircraft propellers is proposed. The feasibility of the method is illustrated by the example of a signal reflected from a Mi-8 helicopter. The requirements to the measurement error of the blade repetition rate and to the frequency analysis step at the precise measurement stage are formulated. The requirement to the repetition rate of probing signals is substantiated, the fulfillment of which ensures an unambiguous restoration of the spectrum of "secondary" modulation of the signal reflected from the blades of a propeller-driven aircraft.Conclusion. The developed method for a two-stage measurement of the blade repetition rate ensures the adaptation of algorithms for constructing radar images of aircraft propellers to their rotation frequency.

Conditional effects of higher order co-moments in asset pricing: Evidence from Borsa Istanbul

This paper explores how systematic higher order moments (co-skewness and co-kurtosis) are priced in Borsa Istanbul. We tested the significance of higher order co-moments and analyzed their contribution to the standard capital asset pricing model and the Fama and French (2015) 5-factor model. We used a two-stage method to analyze the weekly returns of beta and size-sorted portfolios and individual stocks over the sample period from June 22, 2007 to November 15, 2023. We also used models conditional on market movements. The findings reveal that co-skewness has statistically significant effects on portfolio returns in Borsa Istanbul, especially in up markets. We also present the statistically significant effects of co-kurtosis on individual stock returns in both up and down markets.

Study on Mechanical Properties of Sandy Soil Solidified by Enzyme-Induced Calcium Carbonate Precipitation (EICP)

Earth–rock dams are widely distributed in China and play an important role in flood control, water storage, water-level regulation, and water quality improvement. As an emerging seepage control and reinforcement technology in the past few years, enzyme (urease)-induced calcium carbonate precipitation (EICP) has the qualities of durability, environmental friendliness, and great economic efficiency. For EICP-solidified standard sand, this study analyzes the effect of dry density, amount of cementation, standing time, perfusion method, and other factors on the permeability and strength characteristics of solidified sandy soil by conducting a permeability test and an unconfined compression test and then working out the optimal solidification conditions of EICP. Furthermore, a quantitative relationship is established between the permeability coefficient (PC), unconfined compressive strength (UCS), and CaCO3 generation (CG). The test findings indicate that the PC of the solidified sandy soil decreases and the UCS rises as the starting dry density, amount of cementation, and standing time rise. With the increase of CG, the PC of the solidified sandy soil decreases while the UCS increases, indicating a good correlation among PC, UCS, and CG. The optimal condition of solidification by EICP is achieved by the two-stage grouting method with an initial dry density of 1.65 g/cm3, cementation time of 6 d, and standing time of 5 d. Under such conditions, the permeability of the solidified sandy soil is 6.25 × 10−4 cm/s, and the UCS is 1646.94 kPa. The findings of this study are of great theoretical value and scientific significance for guiding the reinforcement of earth–rock dams.

Modified projection generalized two-point two-stage extragradient quasinewton method for saddle point problems

Abstract. The purpose of this work is to investigate a new method mentioned in the article’s name. This method is designed for solving saddle problems with convexo-concave differentiable function that is defined on a convex closed subset of some finite-dimensional euclidean space and has «ravine» level hypersurfaces. The paper contains a brief survey of native publications devoted to new projection gradient methods for solving saddle problems. A mathematical statement of a saddle problem, information about solution method, some auxiliary inequalities, and method’s convergence are discussed in the article as well. Moreover, iterative formulas are exemplified for another perspective saddle method for convexo-concave differentiable saddle functions, which may be validated as well as formulas proved in this work. New auxiliary inequalities complete mathematical apparatus of convex analysis for justification of convergence and rate of convergence and have value also for justification of another methods of operations research. By using obtained inequalities, convex analysis and numerical mathematics, convergence of the saddle method for convexo-concave smooth saddle functions with Lipschitz partial gradients is proved. Under supplementary conditions, for twice continuously differentiable saddle functions, superlinear and quadratic rate of convergence of saddle method are proved, too.

In Ukrainian

Blood loss is one of the main causes of death, especially in war and natural disasters. At the beginning of the full-scale Russian-Ukrainian war we established the laboratory and later the industrial production of a powdered topical hemostatic composition based on nanosilica which is intended for providing first pre-medical aid. The composition contains nanosized silica A-300 and sodium alginate as two active ingredients in a mass ratio of 4:1. The aim of the work was to develop the optimal technological process for the production of a hemostatic composition and carry out physico-chemical and medical-biological studies of semi-finished products and the final product. Bulk density measurements, optical microscopy, IR spectroscopy method and microbiological research were used to study the initial materials, intermediate products as well as the final product. The effectiveness of the hemostatic effect of the composition was checked on the model of parenchymal bleeding from the liver of a rat, using the time to stop bleeding (min) as a criterion. As a result of the research, a two-stage method of manufacturing the composition is proposed: at the first stage, certain parts of the initial materials are treated in a ball mill, obtaining the semi-finished product "A-300/sodium alginate"; at the second stage, this semi-finished product is mixed with nanosilica and sodium alginate, obtaining the final product. It is shown that the bulk density serves as a useful technological parameter to control of which helps to produce a structurally homogeneous final product. In the IR spectra of the semi-finished product and the finished product, only the absorption bands of silica and sodium alginate are observed, that is, foreign substances are not formed during technological process. The microbiological purity of the composition meets the pharmacopoeial requirements for drugs of this category. An experimental study of the topical hemostatic effect of the composition revealed its significant advantage compared to the inorganic hemostatic substance kaolin, which acts according to a similar adsorption mechanism.

Study of Urban Unmanned Aerial Vehicle Separation in Free Flight Based on Track Prediction

In recent years, the application prospect of urban logistics unmanned aerial vehicles has attracted extensive attention. The high-density operation of UAVs requires autonomous separation maintenance capability. To achieve autonomous separation maintenance, it is necessary to conduct autonomous track prediction and formulate the required separation accordingly. Based on the target level of safety requirements for UAV operation, aiming at the autonomous separation maintenance ability of UAVs and considering the accuracy of track prediction, a method to calculate the required separation between UAVs is proposed. This study consists of two parts. Firstly, based on historical data, the position prediction error of the flight track is investigated. Using a machine learning model, a two-stage track prediction method, which involves classification followed by prediction, is proposed for urban logistics UAV track data. Subsequently, based on the track prediction error distribution, by designing a gas model and a position error probability model, a separation-formulating model for urban logistics UAVs in free flight is proposed in which UAV maneuverability is considered. By applying this model, the required separation is formulated for UAVs. When the required separation is set to 48.5 m, the overall collision risk meets the TLS requirements. The research provides a feasible method for establishing autonomous separation for urban logistics UAVs.

An investigation into the role of Residents' cognitive preferences in distributed renewable energy development

During the development and deployment of distributed renewable energy, enhancing residents' cognitive preferences towards renewable energy can effectively mitigate resistance. To examine the effect of residents' cognitive preferences on the development of distributed renewable energy, we first qualitatively and quantitatively analyzed residents' cognitive preferences through literature review and big data analysis. Subsequently, a two-way fixed effects model with time and region were utilized to test the relationship between residents' cognitive preferences and distributed renewable energy. However, the selected sample may lead to biased estimation results due to omitted variable and self-selection bias. To ensure the reliability of the results, we used the instrumental variable method and the Heckman correction method to eliminate these biases. Finally, we used the spatial autoregressive model to explore the spatial interaction relationship between surrounding cognitive activities and residents' cognitive preferences. The results show that residents' cognitive preferences significantly promote the development of distributed renewable energy. Moreover, residents living in areas with higher levels of air pollution, income, education, and informatization are more receptive to the development of distributed renewable energy. Additionally, we found a “herd effect” in residents' cognitive preferences in neighbouring regions, which is limited to adjacent areas. These findings provide an important reference for the government to effectively guide residents' cognition to form the “herd effect”, and then promote the large-scale development of distributed renewable energy.

Illumination-Aware Low-Light Image Enhancement with Transformer and Auto-Knee Curve

Images captured under low-light conditions suffer from several combined degradation factors, including low brightness, low contrast, noise, and color bias. Many learning-based techniques attempt to learn the low-to-clear mapping between low-light and normal-light images. However, they often fall short when applied to low-light images taken in wide-contrast scenes because uneven illumination brings illumination-varying noise and the enhanced images are easily over-saturated in highlight areas. In this article, we present a novel two-stage method to tackle the problem of uneven illumination distribution in low-light images. Under the assumption that noise varies with illumination, we design an illumination-aware transformer network for the first stage of image restoration. In this stage, we introduce the Illumination-aware Attention Block featured with Illumination-aware Multi-head Self-attention, which incorporates different scales of illumination features to guide the attention module, thereby enhancing the denoising and reconstruction capabilities of the restoration network. In the second stage, we innovatively introduce a cubic auto-knee curve transfer with a global parameter predictor to alleviate the over-exposure caused by uneven illumination. We also adopt a white balance correction module to address color bias issues at this stage. Extensive experiments on various benchmarks demonstrate the advantages of our method over state-of-the-art methods qualitatively and quantitatively.

Differences in Associated Factors of Underweight and Overweight According to Rural-Urban Residence Strata among Ever-Married Non-Pregnant Women of Reproductive Age in Bangladesh.

This study sought to compare the prevalence of underweight and overweight among ever-married, non-pregnant women of reproductive age in Bangladesh by urban or rural residency status. This study used Bangladesh Demographic and Health Survey (BDHS), 2017 data. Cross-sectional study design with two-stage stratified sampling method was employed. A sample of ever-married non-pregnant women of reproductive age was selected and multinomial logistic regression was utilised in analysis. It was found that around half of rural women (45.0%, N = 4,934) and more than half of urban women (60.3%, n = 3,913) were overweight. Nearly one in seven rural women (14.0%, n = 1,537) and 1 in 12 urban women (9.0%, n = 564) were reported as underweight. Our analyses revealed that being overweight was substantially connected with age, husband's occupation, economic status, television access, and division for both urban and rural areas. Women from poor households were significantly more likely to be underweight than women from middle- income households for both urban (P < 0.05; OR: 1.41; 95% CI: 1.03, 1.94) and rural (P < 0.05; OR: 1.23; 95% CI: 1.04, 1.46) areas. Interestingly, women without television access both in urban (P < 0.001; OR = 0.78; 95% CI: 0.67, 0.91) and rural (P < 0.001; OR = 0.75; 95% CI: 0.68, 0.84) areas had an inverse association with overweight/obesity compared to women with television access. In both areas, women in Sylhet and Mymensingh had higher likelihood of being underweight than Barisal division. Additionally, in both residential zones, women in Sylhet had lower likelihood of being overweight than Barisal division. This study reveals that multiple characteristics are linked to both overweight and underweight among ever-married, non-pregnant women of reproductive age in Bangladesh. Addressing these variables should be a priority in public health efforts to combat the dual challenge of malnutrition in Bangladesh.

Camouflaged object segmentation with prior via two-stage training

The camouflaged object segmentation (COS) task aims to segment objects visually embedded within the background. Existing models usually rely on prior information as an auxiliary means to identify camouflaged objects. However, low-quality priors and the singular guidance form hinder the effective utilization of prior information. To address these issues, we propose a novel approach for prior generation and guidance, named prior-guided transformer (PGT). For prior generation, we design a prior generation subnetwork consisting of a Transformer backbone and simple convolutions to obtain higher-quality priors at a lower cost. In addition, to fully exploit the backbone’s understanding capabilities of the camouflage characteristics, a novel two-stage training method is proposed to achieve the backbone’s deep supervision. For prior guidance, we design a prior guidance modules (PGM), with distinct space token mixers to respectively capture global dependencies of location priors and local details of boundary priors. Additionally, we introduce a cross-level prior in the form of features to facilitate inter-level communication of backbone features. Extensive experiments have been conducted and experimental results illustrate the effectiveness and superiority of our method. The code is available at https://github.com/Ray3417/PGT.

Microstructure and mechanical properties of copper/al interlayer/graphite joint prepared utilizing “two-stage method” vacuum diffusion bonding

The high property Copper/Graphite joint with Al foil as intermediate layer has been prepared utilizing the “two-stage method” diffusion bonding. At stage I, the Cu–Al diffusion alloying behavior occurs, and when the alloying temperature reaches to 500 °C, the stable Cu9Al4 layer is formed on the surface of Cu substrate. At stage II, Cu (with Cu–Al alloying layer) and Graphite were diffusion bonded, and the microstructure of joint is composed of Cu/Cu9Al4 alloyed layer/Graphite. The C atom diffuses into the Cu9Al4 alloyed layer, which improves the joint property. With the temperature increases from 600 to 900 °C, the diffusion distance of C atom in Cu9Al4 increases from 5.8 μm to 11.4 μm. And the tensile strength of joint increases 25.8 MPa–38.6 MPa. During the diffusion bonding processing, there are three routes for C atom from graphite diffusing in Cu9Al4 alloyed layer, which are OIⅠ→OIⅡ, OIⅠ→TI→OIⅡ and TIⅠ→TIⅡ routes, respectively. The diffusion coefficient of C atoms diffusion along OIⅠ →OIⅡ route is the largest, indicating that OIⅠ→OIⅡ is the preferential diffusion route for C atom diffusing in alloyed layer, and playing the decisive role in the microstructure and properties of joint utilizing two-stage method diffusion bonding.

Determinants of Students’ Satisfaction with AI Tools in Education: A PLS-SEM-ANN Approach

The emergence of Artificial Intelligence (AI) technology has significantly disrupted the educational landscape. The latest development in AI, generative AI that can generate new and tailored to specific content, has significantly impacted education. Given the value of AI technology in general and generative AI specific to users in education, such as students, the adaptability of these technologies has significantly increased. However, continuing and productive usage of AI tools depends upon students’ satisfaction with these tools. Drawing from the existing research, the present research has developed factors that affect students’ general satisfaction with AI tools. The research collected the data using a survey questionnaire from a Saudi Arabian university. The two-stage method of partial least squares structural equation modeling (PLS-SEM) and artificial neural network (ANN) have been employed. The two-stage method is applied in a way that PLS-SEM is used for testing the hypothesis and significance of the factor’s influence on satisfaction, and ANN is used to determine the relevant importance of the factor. The PLS-SEM results have shown that factors such as content quality, emotional wellbeing and perceived utility determine student satisfaction with AI tools. The ANN results show that emotional wellbeing is the most critical factor in satisfaction, followed equally by content quality and perceived utility.

Differential network knockoff filter with application to brain connectivity analysis.

The brain functional connectivity can typically be represented as a brain functional network, where nodes represent regions of interest (ROIs) and edges symbolize their connections. Studying group differences in brain functional connectivity can help identify brain regions and recover the brain functional network linked to neurodegenerative diseases. This process, known as differential network analysis focuses on the differences between estimated precision matrices for two groups. Current methods struggle with individual heterogeneity in measuring the brain connectivity, false discovery rate (FDR) control, and accounting for confounding factors, resulting in biased estimates and diminished power. To address these issues, we present a two-stage FDR-controlled feature selection method for differential network analysis using functional magnetic resonance imaging (fMRI) data. First, we create individual brain connectivity measures using a high-dimensional precision matrix estimation technique. Next, we devise a penalized logistic regression model that employs individual brain connectivity data and integrates a new knockoff filter for FDR control when detecting significant differential edges. Through extensive simulations, we showcase the superiority of our approach compared to other methods. Additionally, we apply our technique to fMRI data to identify differential edges between Alzheimer's disease and control groups. Our results are consistent with prior experimental studies, emphasizing the practical applicability of our method.

Point cloud registration via sampling-based evolutionary multitasking

Point cloud registration (PCR) is an important task in both computer vision and pattern recognition. However, since the success of most existing two-stage PCR methods largely depends on the implicit assumption that the first coarse registration stage can provide an approximately optimal initial transformation for the second fine stage, these methods suffer from low success rate when the assumption is not satisfied. Evolutionary multitasking can find better solutions for multiple tasks simultaneously by transferring knowledge among multiple related tasks. Inspired by evolutionary multitasking, this study proposes a sampling-based multitask optimization method for PCR, which solves the problem of low registration success rate by sharing knowledge between two related tasks generated by a dataset. Specifically, one task is for roughly aligning the sampled point cloud, and the other is for the original unsampled point cloud. The main purpose of the first task is to continuously provide useful knowledge to the second task so that it can achieve precise alignment. Then, a new knowledge transfer mechanism guided by historical information is proposed, which can effectively adjust the search behavior of the original unsampled task. In addition, we develop a dynamic task complement strategy to complement the accuracy that may be lost by downsampling. The results on five different datasets show that the proposed method can achieve higher registration accuracy with higher success rate compared to other state-of-the-art PCR methods.