Explore MATLAB source code curated for "超分辨率重建" with clean implementations, documentation, and examples.
This MATLAB implementation performs image super-resolution reconstruction using the Projection Onto Convex Sets (POCS) method, reconstructing a single high-resolution image from multiple low-resolution input images.
To reduce the computational complexity of motion estimation algorithms and enhance the reliability of sequential image super-resolution reconstruction, we propose an effective three-step search algorithm. This algorithm employs a multi-step search strategy that leverages the center-biased distribution characteristics of motion vectors and parallel processing principles. It replaces conventional square templates with diamond-shaped small templates for refined searching in optimal matching regions, thereby improving search accuracy. Implementation involves coarse-to-fine hierarchical searching with adaptive pattern switching. Experimental results demonstrate significant reduction in computational time while maintaining high search precision.
Implementation of algorithms from the Fast and Robust Multi-Frame Super-Resolution paper, featuring BTV regularization and its enhanced variant for super-resolution reconstruction, including code-level implementation details and algorithmic improvements.
MATLAB implementation of Iterative Back Projection (IBP) algorithm for super-resolution image reconstruction