Average Error Reduction Research Articles

SCGFormer: Semantic Chebyshev Graph Convolution Transformer for 3D Human Pose Estimation

With the rapid advancement of deep learning, 3D human pose estimation has largely freed itself from reliance on manually annotated methods. The effective utilization of joint features has become significant. Utilizing 2D human joint information to predict 3D human skeletons is of paramount importance. Effectively leveraging 2D joint data can improve the accuracy of 3D human skeleton prediction. In this paper, we propose the SCGFormer model to reduce the error in predicting human skeletal poses in three-dimensional space. The network architecture of SCGFormer encompasses Transformer and two distinct types of graph convolution, organized into two interconnected modules: SGraAttention and AcChebGconv. SGraAttention extracts global feature information from each 2D human joint, thereby augmenting local feature learning by integrating prior knowledge of human joint relationships. Simultaneously, AcChebGconv broadens the receptive field for graph structure information and constructs implicit joint relationships to aggregate more valuable adjacent features. SCGraFormer is tested on widely recognized benchmark datasets such as Human3.6M and MPI-INF-3DHP and achieves excellent results. In particular, on Human3.6M, our method achieves the best results in 9 actions (out of a total of 15 actions), with an overall average error reduction of about 1.5 points compared to state-of-the-art methods, demonstrating the excellent performance of SCGFormer.

Research on Inter-Frame Feature Mismatch Removal Method of VSLAM in Dynamic Scenes

Visual Simultaneous Localization and Mapping (VSLAM) estimates the robot’s pose in three-dimensional space by analyzing the depth variations of inter-frame feature points. Inter-frame feature point mismatches can lead to tracking failure, impacting the accuracy of the mobile robot’s self-localization and mapping. This paper proposes a method for removing mismatches of image features in dynamic scenes in visual SLAM. First, the Grid-based Motion Statistics (GMS) method was introduced for fast coarse screening of mismatched image features. Second, an Adaptive Error Threshold RANSAC (ATRANSAC) method, determined by the internal matching rate, was proposed to improve the accuracy of removing mismatched image features in dynamic and static scenes. Third, the GMS-ATRANSAC method was tested for removing mismatched image features, and experimental results showed that GMS-ATRANSAC can remove mismatches of image features on moving objects. It achieved an average error reduction of 29.4% and 32.9% compared to RANSAC and GMS-RANSAC, with a corresponding reduction in error variance of 63.9% and 58.0%, respectively. The processing time was reduced by 78.3% and 38%, respectively. Finally, the effectiveness of inter-frame feature mismatch removal in the initialization thread of ORB-SLAM2 and the tracking thread of ORB-SLAM3 was verified for the proposed algorithm.

ARD-SLAM: Accurate and robust dynamic SLAM using dynamic object identification and improved multi-view geometrical approaches

In the evolving landscape of autonomous navigation, traditional Visual Simultaneous Localization and Mapping (SLAM) systems often encounter challenges in dynamic environments, primarily due to their reliance on assumptions of static surroundings. In response to these limitations, we introduce ARD-SLAM, a groundbreaking approach to dynamic SLAM that innovatively combines global dense optical tracking with sophisticated geometric methodologies. The core innovation of ARD-SLAM lies in its dynamic object identification technique, which harmoniously integrates geometric motion information with prospective motion data. This integration facilitates effective segmentation of moving objects, thereby substantially diminishing their impact on camera ego-motion estimation. ARD-SLAM is further enhanced by an advanced multi-view geometry method that emphasizes the selection of well-matched feature points. This approach is instrumental in efficiently managing dynamic scenarios while also reducing computational load. Rigorous testing on the TUM RGB-D and Bonn RGB-D benchmark datasets has established ARD-SLAM's superiority over established techniques like ORB-SLAM2/3, DynaSLAM, SD-SLAM, DGS-SLAM, and OVD-SLAM. Notably, ARD-SLAM achieves a substantial average reduction in Absolute Trajectory Error (ATE) by 86.1% and in Relative Pose Error (RPE) by 88.0% compared to ORB-SLAM3. The results from the Bonn RGB-D Dataset further underscore ARD-SLAM's effectiveness. Compared to other SLAM methods, ARD-SLAM shows remarkable improvements: 37.8% and 66.4% over DynaSLAM, 41.2% and 73.1% over DGS-SLAM, and 48.9% and 79.7% over OVD-SLAM in ATE and RPE metrics, respectively. This robust performance in dynamically changing environments solidifies ARD-SLAM as a significant advancement in SLAM technology, offering a more precise and adaptable solution for the complex challenges of real-world autonomous navigation.

Short-term PV power prediction based on VMD-CNN-IPSO-LSSVM hybrid model

Abstract This article discusses the significance and obstacles of short-term power prediction in photovoltaic systems and introduces a hybrid model for photovoltaic short-term power prediction technology based on variational mode decomposition (VMD), convolutional neural network (CNN), improved particle swarm optimization (IPSO) and least squares support vector machine (LSSVM). In the initial stage, the photovoltaic generation signal is decomposed into multiple Intrinsic mode functions (IMFs) using VMD to enhance the extraction of signal time–frequency characteristics. Subsequently, CNN is utilized for feature learning and extraction of each IMF, modeling the nonlinear and non-stationary features. Following this, the IPSO-LSSVM optimization algorithm is employed to establish and optimize multiple LSSVM models, predicting power fluctuations at different time scales. Finally, the predictions from each model are synthesized to obtain the final photovoltaic short-term power forecast. Through validation with actual photovoltaic generation data, this hybrid model demonstrates high accuracy and reliability in short-term power prediction, showing an average relative error and root mean square error reduction of 15.23 and 53.60%, respectively, compared to a certain comparative model. This proposed method based on VMD-CNN-IPSO-LSSVM hybrid model for photovoltaic power prediction holds promising prospects and practical value in the operation and scheduling of photovoltaic generation systems.

STICAP : Spatio-temporal Interactive Attention for Citywide Crowd Activity Prediction

Accurate citywide crowd activity prediction (CAP) can enable proactive crowd mobility management and timely responses to urban events, which has become increasingly important for a myriad of smart city planning and management purposes. However, complex correlations across the crowd activities, spatial and temporal urban environment features and their interactive dependencies, and relevant external factors (e.g., weather conditions) make it highly challenging to predict crowd activities accurately in terms of different venue categories (for instance, venues related to dining, services, and residence) and varying degrees (e.g., daytime and nighttime). To address the above concerns, we propose STICAP , a citywide spatio-temporal interactive crowd activity prediction approach. In particular, STICAP takes in the location-based social network check-in data (e.g., from Foursquare/Gowalla) as the model inputs and forecasts the crowd activity within each time step for each venue category. Furthermore, we have integrated multiple levels of temporal discretization to interactively capture the relations with historical data. Then, three parallel Residual Spatial Attention Networks ( RSAN ) in the Spatial Attention Component exploit the hourly, daily, and weekly spatial features of crowd activities, which are further fused and processed by the Temporal Attention Component for interactive CAP . Along with other external factors such as weather conditions and holidays, STICAP adaptively and accurately forecasts the final crowd activities per venue category, enabling potential activity recommendation and other smart city applications. Extensive experimental studies based on three different real-world crowd activity datasets have demonstrated that our proposed STICAP outperforms the baseline and state-of-the-art algorithms in CAP accuracy, with an average error reduction of 35.02%.

A Cubature Kalman Filter for parameter identification and output-feedback attitude control of liquid-propellant satellites considering fuel sloshing effects

The problem of parameter identification and output-feedback attitude stabilization is addressed for liquid-propellant satellites in the presence of fuel sloshing. The sloshing is described by a 2-degree-of-freedom pendulum. A torsional spring-damper system is utilized to model the flexibility of the diaphragm passive strategy and the fuel viscosity. By employing the control Lyapunov function methodology, a nonlinear backstepping-type state-feedback control law is constructed to stabilize the satellite and cancel the fuel sloshing. Then, to make the controller more feasible in practice, the required states of the controller, stiffnesses, and dampers are considered to be unavailable. Thus, a Cubature Kalman Filter is developed to estimate the system states and identify the fuel parameters. The state-feedback control law in conjunction with the Cubature Kalman Filter approach provides an output-feedback control law. This work owns the following two major novelties. One is identifying the stiffness and damping coefficients of the fuel and another is developing a Cubature Kalman Filter for estimating the resultant nonlinear system states. Lastly, comparative simulations verify and highlight the main results of the proposed methodology in relation to the existing methods. The findings suggest that the Cubature Kalman Filter offers potential advantages over the methods outlined in previous works, particularly in terms of reducing RMS error and improving state estimation convergence. The average reduction in RMS error for state convergence is 24 %, while the estimation process is found to be 63 % faster when compared to existing methods.

Research on the Bearing Lifespan Prediction Method for Ship Propulsion Shaft Systems Based on an Enhanced Domain Adversarial Neural Network

To address the challenge of accurate lifespan prediction for bearings in different operating conditions within ship propulsion shaft systems, a two-stage prediction model based on an enhanced domain adversarial neural network (DANN) is proposed. Firstly, pre-training features containing comprehensive degradation information are extracted from the entire source domain dataset encompassing all operational conditions. Subsequently, DANN is employed to extract domain-invariant features that are difficult to distinguish. Following this, a feature alignment process is utilized to align high-dimensional features with pre-training features, thereby mitigating the adverse effects caused by missing data in the incomplete target operational condition dataset. Finally, the effectiveness of this approach is validated using operational data from bearings under multiple operating conditions. The experimental results demonstrate that the method presented in this paper achieves an average error reduction of 0.0626 and 0.0845 compared to the MK-MMD transfer learning method and self-attention ConvLSTM algorithms, respectively, and exhibits higher predictive reliability. This method can provide valuable insights for lifespan prediction challenges concerning bearings in ship propulsion shaft systems under various operational conditions, as well as similar cross-domain lifespan prediction problems.

Optimization of fracture reduction robot controller based on improved sparrow algorithm

The accuracy of a fracture reduction robot (FRR) is critical for ensuring the safety of surgery. Improving the repositioning accuracy of a FRR, reducing the error, and realizing a safer and more stable folding motion is critical. To achieve this, a sparrow search algorithm (SSA) based on the Levy flight operator was proposed in this study for self-tuning the robot controller parameters. An inverse kinematic analysis of the FRR was also performed. The robot dynamics model was established using Simulink, and the inverse dynamics controller for the fracture reduction mechanism was designed using the computed torque control method. Both simulation and physical experiments were also performed. The actual motion trajectory of the actuator drive rod and its error with a desired trajectory was obtained through simulation. An optimized Levy-sparrow search algorithm (Levy-SSA) crack reduction robot controller demonstrated an overall reduction of two orders of magnitude in the reduction error, with an average error reduction of 98.74% compared with the traditional unoptimized controller. The Levy-SSA increased the convergence of the crack reduction robot control system to the optimal solution, improved the accuracy of the motion trajectory, and exhibited important implications for robot controller optimization.

Modeling of gas emission in coal mine excavation workface: a new insight into the prediction model.

Coalbed methane (CBM) is a clean energy source, but its utilization is inefficient due to the complexity and low accuracy of its emission prediction model. In this research, we constructed a mathematical model of gas emission from the excavation workface, and combined the experimental results to propose a new model for accurate and concise prediction. The new model was validated in the field workface and compared with the traditional prediction model. Moreover, the sensitivity of gas emission parameters and the participation ratio of gas emission sources were analyzed. The study results show that the new model has higher calculation accuracy than the old model prediction, with an average error reduction of 4.693%. In the excavation workface, the coal fall gas emission conforms to the negative power function equation, and the coal wall gas emission conforms to the negative exponential function equation. In the early stage of excavation, the proportion of coal fall gas emission is higher than that of coal wall gas emission, and the peak proportion reaches 58.5%. In the later stage, the proportion of coal fall gas emission gradually decreases to below 30%. The order of the sensitivity of gas emission parameters is coal wall gas initial velocity > coal fall gas decay coefficient > coal fall gas initial velocity > coal wall gas decay coefficient. The new model is successfully applied in engineering, which helps to improve the efficiency of coal mine gas disaster control and utilization.

Optimal Discrete Element Parameters for Black Soil Based on Multi-Objective Total Evaluation Normalized-Response Surface Method

The lack of accurate black soil simulation model parameters in the design and optimization of soil remediation equipment has led to large errors in simulation results and simulation outcomes, which to some extent restricts the development of soil remediation equipment. Accurate discrete element parameters can improve the efficiency of soil remediation equipment. To improve the reliability of the discrete element contact parameters for black soil, a set of optimal discrete element contact parameters was found that could comprehensively represent a variety of particle sizes and minimize error. In this paper, the best discrete element contact parameters were selected by using a multi-indicator total evaluation normalization method combined with the response surface method, combined with black soil solid and simulated stacking tests. First, the physical parameters of the black soil and the accumulation angle were determined. Next, Plackett–Burman tests were carried out for each grain size in turn to obtain the contact parameters that had a significant effect on the black soil accumulation angle. The important parameters obtained for different particle sizes are all as follows: black soil–black soil static friction coefficient, black soil–black soil rolling friction coefficient, and black soil–stainless steel rolling friction coefficient. In conjunction with the Plackett–Burman test screening results, the steepest climb test was designed for six grain sizes to optimize the range of values. To find the optimal contact parameters for the different particle sizes based on the final results of Box–Behnken experiments, the discrete element parameters of black soil were optimized for the different particle sizes of black soil by using the multi-indicator total evaluation normalization method and response surface method. The results showed that the black soil–black soil static friction coefficient was 1.045, the black soil–black soil rolling friction coefficient was 0.464, and the black soil–stainless steel rolling friction coefficient was 0.215. The errors for each particle size were reduced by 0.89%, 0.7%, 0.84%, 0.57%, 0.71%, and 0.76% for the best combination of parameters before and after normalization, with an average error reduction of 0.745%. This data provides some reference value for the design and optimization of soil remediation equipment.

Event fusion photometric stereo network

Photometric stereo methods typically rely on RGB cameras and are usually performed in a dark room to avoid ambient illumination. Ambient illumination poses a great challenge in photometric stereo due to the restricted dynamic range of the RGB cameras. To address this limitation, we present a novel method, namely Event Fusion Photometric Stereo Network (EFPS-Net), which estimates the surface normals of an object in an ambient light environment by utilizing a deep fusion of RGB and event cameras. The high dynamic range of event cameras provides a broader perspective of light representations that RGB cameras cannot provide. Specifically, we propose an event interpolation method to obtain ample light information, which enables precise estimation of the surface normals of an object. By using RGB-event fused observation maps, our EFPS-Net outperforms previous state-of-the-art methods that depend only on RGB frames, resulting in a 7.94% reduction in mean average error. In addition, we curate a novel photometric stereo dataset by capturing objects with RGB and event cameras under numerous ambient light environments.

A New Approach for Predicting the Rheological Properties of Oil-Based Drilling Fluids under High Temperature and High Pressure Based on a Parameter-Free Method

Under different temperatures and pressures, the physical parameters of drilling fluid will change, resulting in inaccurate drilling hydraulic calculations. Aiming to address the problem of the traditional rheological prediction method needing to first determine the rheological model, this paper proposed a method for first predicting the readings of the rheometer and then determining the rheological model. The model established in this paper adopted a parameter-free method, which expands the application range of the model. Rheology experiments were carried out on the three types of oil-based drilling fluids collected at the well site. The model in this paper was verified based on the experimental data. The results showed that, compared with the traditional drilling fluid rheological prediction method, the model established in this paper had a better prediction effect, with an average error of 4.85%, and the average error reduction ranges from 3.8% to 8.3%. The model established in this paper is able to provide theoretical support for accurate hydraulic calculation.

Effects of Error Modulation-Based Visual and Haptic Feedback Fusion Strategies on Motor Learning and Motivation

Visual or haptic feedback based on error modulation has been used to improve the effect of robot-assisted rehabilitation training. However, there are several investigations on the effects of error modulation-based visual and haptic feedback fusion strategies on motor learning and motivation. To observe the influence of different feedback fusion strategies on motor learning and motivation, a parallel controlled study was conducted, dividing 30 healthy subjects into three groups with similar skill levels. The no error modulation group received visual and haptic feedback without error modulation; the visual amplification haptic reduction group received visual error amplification combined with haptic error reduction, and the visual reduction haptic amplification (VRHA) group received visual error reduction combined with haptic error amplification. Each subject implemented a trajectory-tracking task with an upper limb rehabilitation robot. They went through baseline, training, assessment, and generalization tests and completed 340 consecutive tracking movements. To evaluate motor learning and motivation, the average tracking error, the root mean square (RMS) of surface electromyography (sEMG) signals, and the intrinsic motivation inventory scale were all examined. In the assessment tests, the average tracking error was significantly decreased in all three groups. In particular, the VRHA group had a larger reduction in average tracking error in the generalization test, lower RMS of sEMG signals both in the assessment and generalization tests and higher perceived competence in the assessment tests. The VRHA fusion strategy significantly improved the subjects’ motor learning and transfer ability, decreased muscle activation, and increased motor learning motivation. These findings may provide some new insights for multisensory feedback fusion technology in the application of rehabilitation robots.

Time-series generative adversarial networks for flood forecasting

Reliable flood forecasting serves as one of the fundamental tasks in flood management. However, the forecast performance of modern data-driven models depends heavily on the quantity and quality of training data. The lack of field data of flood events highly underpins the development of machine learning (ML) flood forecasting techniques. With consideration of the rareness of flood events and the high dimensionality of flood time series, two latest variants of generative adversarial networks (GANs), Time-series Generative Adversarial Network (TimeGAN) and Real-world Time Series Generative Adversarial Network (RTSGAN), were applied to generate synthetic flood timeseries. In the example analysis of the Xijiang River Basin in southern China, results show that time-series GANs can accurately and efficiently mimic the spatiotemporal correlations between flood series from multiple sites, and RTSGAN outperforms TimeGAN especially when the lengths of time series are long. Additionally, gradient boosting regression tree (GBRT), long-short term memory (LSTM) and quantile regression method integrated LSTM networks (QRLSTM) flood forecasting models were trained with real sequences and an extra amount of synthetic sequences. It shows that the expansion of training datasets in this way is promising in reducing prediction errors of classical machine learning models. In the 3-day-ahead flood forecast, the introduction of synthetic training datasets has little positive effect on both GBRT and LSTM. In 24-hour-ahead flood forecast, as the lead period increases, GBRT-RTSGAN outperforms GBRT evidenced by a 6.7% increase in Nash-Sutcliffe efficiency (NSE) metric, a 56.5% reduction in average absolute relative error (AARE) metric and a 33.1% reduction in root-mean square error (RMSE) metric. For interval forecast, the use of synthetic training datasets helps QRLSTM obtain lower prediction interval normalized averaged width (PINAW), however, at the cost of prediction interval coverage probability (PICP).

Modelling the barotropic sea level in the Mediterranean Sea using data assimilation

Abstract. This paper analyses the variability of the sea level barotropic components in the Mediterranean Sea and their reproduction using a hydrodynamic model with and without data assimilation. The impact of data assimilation is considered both in reanalysis and short-forecast simulations. We used a two-dimensional finite element model paired with an ensemble Kalman filter, which assimilated hourly sea level data from 50 stations in the Mediterranean basin. The results brought about a significant improvement given by data assimilation in the reanalysis of the astronomical tide, the surge, and the barotropic total sea level, even in coastal areas and far from the assimilated stations (e.g. the southeastern Mediterranean Sea). As with the reanalysis simulations, the forecast simulations, which start from analysis states, improve, especially on the first day (37 % average error reduction) and when seiche oscillations are triggered. Since seiches are free barotropic oscillations that depend only on the initial state, their reproduction improves very effectively with data assimilation. Finally, we estimate the periods and the energy of these oscillations by means of spectral analysis, both in the Adriatic Sea, where they have been extensively studied, and in the Mediterranean Sea, where the present documentation is scarce. While the periods are well reproduced by the model even without data assimilation, their energy shows a good improvement when using it.

Wi-Breath: A WiFi-Based Contactless and Real-Time Respiration Monitoring Scheme for Remote Healthcare.

Respiration rate is an important healthcare indicator, and it has become a popular research topic in remote healthcare applications with Internet of Things. Existing respiration monitoring systems have limitations in terms of convenience, comfort, and privacy, etc. This paper presents a contactless and real-time respiration monitoring system, the so-called Wi-Breath, based on off-the-shelf WiFi devices. The system monitors respiration with both the amplitude and phase difference of the WiFi channel state information (CSI), which is sensitive to human body micro movement. The phase information of the CSI signal is considered and both the amplitude and phase difference are used. For better respiration detection accuracy, a signal selection method is proposed to select an appropriate signal from the amplitude and phase difference based on a support vector machine (SVM) algorithm. Experimental results demonstrate that the Wi-Breath achieves an accuracy of 91.2% for respiration detection, and has a 17.0% reduction in average error in comparison with state-of-the-art counterparts.

A study on application of machine-learning on DBI soot diagnostics

Nowadays, Diffused Background-illumination Extinction Imaging technique (DBI) has been widely applied to quantify the in-flame soot formation of diesel-like sprays. In order to eliminate the errors on soot KL value brought from flame radiation, the soot radiation images are usually needed to be recorded between every two successive back light-on pulses. Consequently, it is necessary to do the frame interpolation for the missing flame radiation images and back light-on images. In this study, a Super SloMo machine-learning method was applied to generate the missing frames and the accuracy of interpolated frames was evaluated from three aspects, namely the frame interval length, prediction region and number of intermediate frames. Meanwhile, Gunner Farnebäck method was applied here as a reference. The results show that the Super SloMo method consistently outperformed the Gunner Farnebäck method in terms of accuracy, regardless of the variations on frame intervals, soot regions, as well as interpolated frame numbers. Consequently, the application of the Super SloMo method in conjunction with the DBI technique significantly improved the accuracy of soot KL value, resulting in a 26.4% reduction in average KL error compared to the Gunner Farnebäck method in the studied cases.

A Novel Method for Imputing Missing Values in Ship Static Data Based on Generative Adversarial Networks

Ship data obtained through the maritime sector will inevitably have missing values and outliers, which will adversely affect the subsequent study. Many existing methods for missing data imputation cannot meet the requirements of ship data quality, especially in cases of high missing rates. In this paper, a missing data imputation method based on generative adversarial networks (GANs) is proposed. The generative adversarial imputation network (GAIN) is improved using the Wasserstein distance and gradient penalty to handle missing values. Meanwhile, the data preprocessing process is optimized by combining knowledge from the ship domain, such as using isolation forests for anomaly detection. Statistical analysis of ship data is also conducted, including correlation analysis of ship design parameters, analysis of outliers, and analysis of missing data types. These analyses provide the basis for the proposed model. In a case study of 8167 bulk carriers, the proposed model outperformed the missing forest (MF) and polynomial fitting (PF) models, with an average error reduction of 2.4% and 6.3%, respectively. The proposed model also showed stable performance in cases of high missing rates. This study provides a new approach for estimating or imputing critical parameters of ships.

Improved LeGO-LOAM method based on outlier points elimination

Based on the analysis of the shortcomings of Lightweight and Ground-Optimized Lidar Odometry and Mapping (LeGO-LOAM) algorithm, an improved LeGO-LOAM method based on outlier points elimination (LeGO-LOAM-OE) was proposed. The original point cloud data generated by Light Detection and Ranging (LiDAR) was analyzed and the four types of outlier points were eliminated, which made the extracted feature points more effective and improved the accuracy of pose estimation. In ground segmentation, the levelness of the ground is constrained so that the extracted ground points are more abundant and accurate. In dynamic point elimination, dynamic objects are eliminated by the moving distance of the same cluster objects in adjacent frames, and the calculation method is simple. The simulation results show that the trajectory accuracy generated by LeGO-LOAM-OE is higher than that generated by LeGO-LOAM. Compared with LeGO-LOAM, LeGO-LOAM-OE can effectively reduce the maximum error and average error of three axes. The total maximum error is reduced by 14.3%, and the total mean error is reduced by 7.6%. LeGO-LOAM-OE had the most significant effect on reducing the Y-axis error (vertical direction), with an average error reduction of 8.5%. An experimental platform was built to test the positioning accuracy and mapping effect of the proposed algorithm in an outdoor environment. Experimental results showed that the proposed method has high accuracy in both online and offline simultaneous localization and mapping (SLAM). The maximum error of the calculated position is less than 1 m, and the average error is less than 0.1 m. Moreover, the generated point cloud map can highly reproduce the environment without deviation.

Break and Splice: A Statistical Method for Non‐Rigid Point Cloud Registration

Abstract3D object matching and registration on point clouds are widely used in computer vision. However, most existing point cloud registration methods have limitations in handling non‐rigid point sets or topology changes (e.g. connections and separations). As a result, critical characteristics such as large inter‐frame motions of the point clouds may not be accurately captured. This paper proposes a statistical algorithm for non‐rigid point sets registration, addressing the challenge of handling topology changes without the need to estimate correspondence. The algorithm uses a novel Break and Splice framework to treat the non‐rigid registration challenges as a reproduction process and a Dirichlet Process Gaussian Mixture Model (DPGMM) to cluster a pair of point sets. Labels are assigned to the source point set with an iterative classification procedure, and the source is registered to the target with the same labels using the Bayesian Coherent Point Drift (BCPD) method. The results demonstrate that the proposed approach achieves lower registration errors and efficiently registers point sets undergoing topology changes and large inter‐frame motions. The proposed approach is evaluated on several data sets using various qualitative and quantitative metrics. The results demonstrate that the Break and Splice framework outperforms state‐of‐the‐art methods, achieving an average error reduction of about 60% and a registration time reduction of about 57.8%.