Abstract:During matching on feature point, gray correlation matching technology is utilized to extract multi-peaks as a coarse matching set. A pair of given corresponding reference points within the left and right images is used to calculate gradients of reference difference between the reference points and each feature point within the multi-peaks set. The unique correspondence is determined by criterion of minimal gradients of reference difference. The obtained correspondence is taken as a new pair of reference points to update the reference points continuously until all feature points in the left (or right) image being matched with the right (or left) image. The gradients of reference difference can be calculated easily by means of pre-setting a pair of obvious feature points in the left and right images as a pair of corresponding reference points. Besides, the efficiency of matching can be improved greatly by taking the obtained matching point as a new pair of reference points, and by updating the reference point continuously. It is proved that the proposed algorithm is valid and reliable by 3D reconstruction on two pairs of actual natural images with abundant and weak texture, respectively.

Point Pattern Matching Based on Spectra of Directed Graphs

Point Pattern Matching: Spectral Descriptor Based Approach

Point Pattern Matching Based on Multiple Spectral Representations

Point Pattern Matching Based on Line Graph Spectral Context and Descriptor Embedding.

A Point Pattern Matching Algorithm Based on Local Relative Shape Context and Q-Spectra

Point Pattern Matching Algorithm Based on Q-Spectrum of Line Graph

A Point Pattern Matching Algorithm Based on Minimize Spanning Tree and Fiedler Vector

Method with the center of graph for point pattern matching

Feature Matching Based on Laplacian Spectra of Graphs

Robust point pattern matching based on spectral context

Gradient of Reference Difference Based Matching Algorithm for Image Feature Point

A Practical Fast Point Pattern Matching Algorithm Based On Matching Pair Support And Registration Condition

Improved Spectral Method for Point Pattern Matching

Feature Points Matching Based on Angle between Spectral Coefficient of Images

"Hedgehog Point" Feature Point Matching Based on Point Structure Information

Inexact Point Pattern Matching Algorithm Based on Delaunay Triangularization and Spectral Method

An Image Matching Algorithm Based on Spectral Features

Spectral Feature Matching Based on Isometric Projection of Matrix

An Algorithm For Point Correspondences Based On Laplacian Spectra Of Graphs

Point Pattern Matching Algorithm Based on Relative Shape Context and Spectral Matching Method

A Geometric Point Sets Pattern Matching by Motion Estimation