Abstract:Learning based super-resolution (SR) methods, which predict the high-resolution pixel values but not directly provide an estimation of uncertainty, are typically non-probabilistic and have limited generalization ability. Gaussian processes can provide a framework for deriving regression techniques with explicit uncertainty models, but Gaussian Process Regression (GPR) has a significant drawback in being time consuming. The computational complexity of GPR is cubic in the number of training examples, which is prohibitively expensive for a large-scale training set. In this article, we have proposed learning local GPR for image SR. Two algorithms are developed to support local GPR for super resolution. A data-driven GPR based super-resolution algorithm is first developed to learn a local GPR model for every LR patch on an input oriented training dataset with moderate size. In order to further improve the running speed, a prototype based GPR algorithm is developed for super resolution. The proposed algorithm is about one-order faster than the data-driven GPR solution because it makes models for the prototypes of image patches rather than for each image patch. Thus, the local regression efforts are greatly reduced to just finding the nearest prototype for each LR image patch and applying its corresponding pre-computed projective matrix for super-resolution prediction. Our algorithms have greater robustness and usability as they provide a formularized way to automatically learn the hyper-parameters introduced for optimizing the covariance function, while most of the state-of-the-art super-resolution methods could only utilize these parameters in a cross-verification way. Moreover, our algorithms offer confidence values at the test points which benefit the pixels' post-processing. Our algorithms are evaluated on popular datasets that are widely used in the super-resolution literature, and the experimental results have demonstrated that the efficiency and effectiveness of our proposed algorithms are comparative with several state-of-the-arts super-resolution methods.

A Gaussian process guided super resolution sampling strategy for the efficient recovery of complex surfaces

A Data-Driven Machining Errors Recovery Method for Complex Surfaces with Limited Measurement Points

Domain-specific Gaussian Process-Based Intelligent Sampling for Inspection Planning of Complex Surfaces

GP-SLAM plus : real-time 3D lidar SLAM based on improved regionalized Gaussian process map reconstruction

Space-filling Scan Paths and Gaussian Process-Aided Adaptive Sampling for Efficient Surface Measurements

Gaussian Process Based Intelligent Sampling for Measuring Nano-Structure Surfaces

An Adaptive Sampling Approach for Digitizing Unknown Free-form Surfaces Based on Advanced Path Detecting

Enhancing the Metrological Performance of Non-Raster Scanning Probe Microscopy Using Gaussian Process Regression

Fast Single Image Super-Resolution Using Sparse Gaussian Process Regression

Feature Sensitive Re-Sampling of Point Set Surfaces with Gaussian Spheres

GausSurf: Geometry-Guided 3D Gaussian Splatting for Surface Reconstruction

Generative Model-Driven Sampling Strategy for the High-Efficiency Measurement of Complex Surfaces on Coordinate Measuring Machines.

High-quality Surface Reconstruction using Gaussian Surfels

DynaSurfGS: Dynamic Surface Reconstruction with Planar-based Gaussian Splatting

A Sparse-Point Guided Photometric Stereo Method for the Metal Complex Surfaces Measurement and High-Fidelity Geometry Recovery

3DGSR: Implicit Surface Reconstruction with 3D Gaussian Splatting

GSurf: 3D Reconstruction via Signed Distance Fields with Direct Gaussian Supervision

StreetSurfGS: Scalable Urban Street Surface Reconstruction with Planar-based Gaussian Splatting

Geodesic Gaussian Processes for the Parametric Reconstruction of a Free-Form Surface

Learning Local Gaussian Process Regression for Image Super-Resolution.

Dependant Gaussian Processes Regression for Intelligent Sampling of Freeform and Structured Surfaces