Abstract:Surface registration, the task of aligning several multidimensional point sets, is a necessary task in many scientific fields. In this work, a novel statistical approach is developed to solve the problem of nonrigid registration. While the application of an affine transformation results in rigid registration, using a general nonlinear function to achieve nonrigid registration is necessary when the point sets require deformations that change over space. The use of a local likelihood-based approach using windowed Gaussian processes provides a flexible way to accurately estimate the nonrigid deformation. This strategy also makes registration of massive data sets feasible by splitting the data into many subsets. The estimation results yield spatially-varying local rigid registration parameters. Gaussian process surface models are then fit to the parameter fields, allowing prediction of the transformation parameters at unestimated locations, specifically at observation locations in the unregistered data set. Applying these transformations results in a global, nonrigid registration. A penalty on the transformation parameters is included in the likelihood objective function. Combined with smoothing of the local estimates from the surface models, the nonrigid registration model can prevent the problem of overfitting. The efficacy of the nonrigid registration method is tested in two simulation studies, varying the number of windows and number of points, as well as the type of deformation. The nonrigid method is applied to a pair of massive remote sensing elevation data sets exhibiting complex geological terrain, with improved accuracy and uncertainty quantification in a cross validation study versus two rigid registration methods.

Estimates of Non-Stationary Spatial Structure Using Gaussian Radial Basis Deformations

Modelling of Non-Stationary Spatial Structure Using Parametric Radial Basis Deformations

Local Linear M‐estimation in Non‐parametric Spatial Regression

A Generalized Spatial Autoregressive Neural Network Method for Three-Dimensional Spatial Interpolation

Estimation of Spatial Deformation for Nonstationary Processes via Variogram Alignment

Flexible Basis Representations for Modeling Large Non-Gaussian Spatial Data

Reduced Basis Kriging for Big Spatial Fields

Spatial Deep Convolutional Neural Networks

deform: An R Package for Nonstationary Spatial Gaussian Process Models by Deformations and Dimension Expansion

A Real-Time Spatial Deformation Estimation Method Based on Spatial Curvature Decomposition and Interpolation

Detecting Statistically Significant Geographical Anomalous Regions from Spatial Sampling Points by Coupling Gaussian Function and Multidirectional Optimization.

Deformation Field Formation Algorithm Based on Modified Kriging Interpolator in GNSS-Based InBSAR

Nonrigid registration using Gaussian processes and local likelihood estimation

Gaussian Process Regression for Transportation System Estimation and Prediction Problems: the Deformation and a Hat Kernel

Nonparametric Density Estimation for Data Scattered on Irregular Spatial Domains: A Likelihood-Based Approach Using Bivariate Penalized Spline Smoothing

Effects of the Factors for Spatial Correlation Analysis on Kriging Estimation in Reservoir Modeling

Estimating Basis Functions in Massive Fields under the Spatial Mixed Effects Model

Soft Tissue Deformation Estimation Model Based on Spatial-Temporal Kriging for Needle Insertion Procedure

Gaussian process regression-based forecasting model of dam deformation

Combining Regression Kriging With Machine Learning Mapping for Spatial Variable Estimation

Spatio-temporal DeepKriging for Interpolation and Probabilistic Forecasting