Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A local visual feature selection method and device

A visual feature and feature selection technology, applied in the computer field, to achieve the effect of reliable retrieval results

Active Publication Date: 2019-05-28
PEKING UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] Aiming at the defects of the prior art, the present invention provides a local visual feature selection method and device, which can solve the problem that the local visual feature selection method in the prior art cannot make the selected partial visual feature subset include as many as possible located on the query target. The problem of local visual features of

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A local visual feature selection method and device
  • A local visual feature selection method and device
  • A local visual feature selection method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0055] figure 1 is a schematic flow chart of a local visual feature selection method provided in Embodiment 1 of the present invention, as shown in figure 1 As shown, the method includes the following steps:

[0056] S101: Detect and acquire multiple local visual features in the target image, and obtain the own attributes of each local visual feature.

[0057] Specifically, the local visual feature detection method is used to detect the local visual features in the target image, form a complete set of local visual feature descriptors, and obtain the own attributes of each local visual feature.

[0058] Wherein, the local visual feature detection method may be any detection method suitable for local visual features of an image, such as a local visual feature detection method in Scale Invariant Feature Transform (SIFT for short).

[0059] Wherein, the own attributes of the local visual features may be the attributes contained in the local visual features detected by the above ...

Embodiment 2

[0087] image 3 is a schematic flow chart of a local visual feature selection method provided in Embodiment 2 of the present invention, as shown in image 3 As shown, the method includes the following steps:

[0088] S301: Detect and acquire multiple local visual features in the target image, and obtain the own attributes of each local visual feature.

[0089] This step S301 is the same as the step S101 in Embodiment 1, and will not be repeated here.

[0090] S302: Obtain depth information of all the local visual features, and obtain depth attributes of the local visual features through a normalization method.

[0091] Optionally, for a local visual feature, the acquired depth information may be a depth value of the local visual feature from the camera, or a disparity value of the local visual feature.

[0092] The local visual features of the target image are all concentrated. In order to obtain the depth value of each local visual feature from the camera, the method of ob...

Embodiment 3

[0113] Figure 4 It is a schematic flowchart of a local visual feature selection method provided by Embodiment 3 of the present invention. The difference between this embodiment and Embodiment 2 is that some depth information acquisition methods can only obtain a part of the complete set of local visual features in the target image. Depth information of visual features. In this case, this embodiment provides another possible numerical calculation method for local visual features.

[0114] Specifically, such as Figure 4 As shown, the method includes the following steps:

[0115] S401: Detect and acquire multiple local visual features in the target image, and obtain the own attributes of each local visual feature.

[0116] This step S401 is the same as the step S101 in Embodiment 1, and will not be repeated here.

[0117] S402: Obtain depth information of some of the partial visual features, and obtain depth attributes of the partial visual features through a normalization m...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The present invention provides a local visual feature selection method and device, the method comprising: detecting and acquiring multiple local visual features in a target image, and obtaining the own attributes of each local visual feature; obtaining part or all of the local visual features Depth attributes of visual features; according to the own attributes of each local visual feature, and the depth attributes of some or all of the local visual features, using a pre-established feature selection model to obtain each of the local visual features located in the query target Possibility values: selecting a preset number of local visual features as a subset of local visual features in descending order of the probability values. The present invention enables the selected local visual features to contain as many local visual features as possible in the query target under the configuration with a small number of features in the partial visual feature subset, while maintaining the compact expression of the visual feature descriptor Under the premise of ensuring more reliable search results.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a local visual feature selection method and device. Background technique [0002] With the popularity of smart mobile devices and the mobile Internet, mobile visual search applications are increasing. Typically, image retrieval relies on local visual features for query and matching. A typical image retrieval system architecture is: the client extracts the local visual features of the query image, and sends the local visual feature descriptors to the server; the server combines the received local visual feature descriptors with a large number of images stored in advance The local visual feature descriptors are matched, and the image with the highest similarity of the local visual feature descriptors is returned as the query result. Due to the memory and bandwidth limitations of mobile smart devices and mobile Internet, it is often necessary to compactly express local d...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06F16/583G06K9/46G06K9/62
CPCG06F16/5838G06V10/462G06F18/22
Inventor 段凌宇刘赵梁陈杰黄铁军高文
Owner PEKING UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com