Method and a device for image matching

Patent metadata

Field	Value
Publication number	US-10699156-B2
Application number	US-201615749545-A
Country	US
Kind code	B2
Filing date	Jan 13, 2016
Priority date	Jan 13, 2016
Publication date	Jun 30, 2020
Grant date	Jun 30, 2020

Abstract

A method for image matching includes acquiring a template image and a target image; acquiring a group of template features according to the template image; extracting a group of target features according to the target image; and according to template features and target features, calculating an degree of image similarity between the template image and each frame of target images, and using a target image with the maximum degree of image similarity as a matched image to the template image. In the image-matching method, image matching is performed by calculating an degree of image similarity between a template image and each target image according to a degree of image similarity between template features and target features, so that non-redundancy of features in an image matching process and correct image matching can be guaranteed, and the image matching accuracy can be improved.

First claim

What is claimed is:

1. A method for image matching, comprising:
acquiring a frame of a template image;
acquiring a plurality of frames of target images;
acquiring a group of template features according to the template image, wherein the group of template features comprise a plurality of template features;
extracting a group of target features from each frame of target images based on pixel gray scale features of each frame of target images, wherein the group of target features comprises a plurality of target features;
according to the group of template features of the template image and a group of target features in each frame of the target images, calculating a degree of image similarity between the template image and each frame of target images to obtain a plurality of degrees of image similarity, comprising:
on each frame of target images, acquiring the inter-feature similarity between each of the group of template features of the template image and each of the group of target features of a current target image;
searching for matched target features which match each template feature in the group of template features of the template image from the group of target features of the current target image according to the inter-feature similarity; and
calculating the degree of image similarity between the current target image and the template image according to the inter-feature similarity between the group of template features and the matched target features of the current target image,
wherein searching for matched target features comprises:
selecting a plurality of target features that are closest to the current template feature from the group of target features of the current target image according to the inter-feature similarity between the current template feature and each of the group of target features of the current target image, wherein the plurality of target features are only a portion of the group of target features of the current target image;
calculating a weighted average of the plurality of inter-feature similarities between the current template features and the plurality of target features; and
acquiring the inter-feature similarity closest to the weighted average from the plurality of inter-feature similarities, and using the target feature corresponding to the inter-feature similarity closest to the weighted average as a matched target feature matching the current template feature;

acquiring a maximum degree of image similarity according to the plurality of degrees of image similarity;
acquiring a target image having maximum degree of image similarity; and
using the target image corresponding to the maximum degree of image similarity as a matched image to the template image.

2. The method of claim 1 , wherein acquiring a group of template features according to the target image comprises:
receiving a user's input and acquiring a group of template features according to the user's input; or
extracting the group of template features from the template image according to the pixel gray scale features of the template image.

3. The method of claim 1 , wherein acquiring a group of target features from each frame of target images comprises:
extracting the group of target features from one of target images using a method from a group comprising Maximally Stable Extremal Regions (MSER), Scale Invariant Feature Transform (SIFT), Hessian, Harris Affine, or Histogram Attribute Relational Graph (HARG).

4. The method of claim 1 , wherein the step of selecting a plurality of target features that are closest to the current template feature from the group of target features of the current target image comprises:
selecting three target features closest to the current template feature from the target features of the group of target features of the current target image.

5. The method of claim 1 , wherein calculating the degree of image similarity between the current target image and the template image according to the inter-feature similarity between the group of template features and the matched target features of the current target image comprises:
calculating a sum of the degrees of image similarity between the group of template features and the matched target features of the current target image as the degree of image similarity between the current target image and the template image.

6. The method of claim 1 , wherein the template feature is a template feature point, the target feature is a target feature point, the inter-feature similarity is a degree of image similarity between feature points, and the matched target feature is a matched target feature point,
wherein the template feature is a template feature edge, the target feature is a target feature edge, the inter-feature similarity is a degree of image similarity between feature edges, and the matched target feature is a matched target feature edge.

Assignees

• Univ Peking Shenzhen Graduate School

Inventors

• Wang Wenmin
• Zhang Ruonan
• Wang Ronggang
• Li Ge
• Dong Shengfu
• Wang Zhenyu
• Li Ying
• Gao Wen

Classifications

• G06V10/945

  User interactive design; Environments; Toolboxes · CPC title

• G06V10/771

  Feature selection, e.g. selecting representative features from a multi-dimensional feature space · CPC title

• G06V10/462 Primary

  Salient features, e.g. scale invariant feature transforms [SIFT] · CPC title

• G06V10/757

  Matching configurations of points or features · CPC title

• G06F18/21347

  using domain transformations · CPC title