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

Cloud image recognition method based on vocabulary tree retrieval and similarity verification

An image recognition and similarity technology, applied in the field of image recognition, can solve problems such as the speed of uploading images, and achieve the effect of improving retrieval speed, fast retrieval, and reducing requirements

Inactive Publication Date: 2017-06-20
成都弥知科技有限公司
View PDF3 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] However, the current cloud-based real-time image recognition technology uploads local pictures to the server, and the server compares the pictures with recognition with the stored pictures one by one. It has the following defects: in the case of poor wireless networks, users can upload images in real time. The speed of the

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

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0039] This embodiment discloses a method for generating a cloud image database based on description sub-samples, including the following steps:

[0040] Descriptor generation step: collect pictures, extract ORB feature points of each picture, and generate corresponding descriptors for each ORB feature point to obtain descriptor samples;

[0041] Step of generating a tree model: generating a tree model of the image database according to the description sub-sample;

[0042] Database generation step: adding pictures to the tree model to establish a tree structure image database.

[0043] specific:

[0044] In the step of generating descriptors, the number of pictures collected should be many and come from various scenes, generally tens of thousands of pictures are required, which are stored in a folder, and commonly used picture formats are acceptable, such as JPG, JPEG, JPE, JFIF, BMP ; Each image is scaled to a certain scale to establish an image pyramid, and the ORB algorit...

Embodiment 2

[0057] A cloud image recognition method based on vocabulary tree retrieval and similarity verification, comprising the following steps,

[0058] Image acquisition step: acquire the target image, and use the ORB algorithm to extract all ORB feature points on the target image, and generate a corresponding descriptor for each ORB feature point, and generate an ORB descriptor sequence of the target image;

[0059] Image uploading step: upload the ORB description subsequence to the cloud image database based on the description sub-samples;

[0060] Image recognition step: the cloud image database uses a vocabulary tree-based retrieval algorithm to match and identify images and return N candidate images with the highest matching degree, where N is a natural number greater than 1;

[0061] Similarity verification step: find the candidate images in the cloud image database, get the 128-dimensional vectors of the target image and the candidate images, calculate the distance between the...

Embodiment 3

[0066] In Embodiment 2, obtaining the 128-dimensional vectors of the target image and the candidate image is performed in the similarity verification system. This embodiment refines the generation method of the similarity verification system.

[0067] This embodiment is based on the classic ImageNet image library and neural network model on the network, which is carried out in the embedded system. Of course, other image libraries may also be utilized. It includes the following steps,

[0068] C1. Input the images in the image library into the neural network model to obtain the 1024-dimensional normalized descriptor corresponding to each image;

[0069] C2. Carry out three-byte learning on the image in the image library, that is, each three-byte contains two positive samples and one negative sample, so as to establish a close distance between positive samples and positive samples, positive samples and negative samples distance between them.

[0070] At this point, if any im...

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 invention discloses a cloud image recognition method based on vocabulary tree retrieval and similarity verification. The method includes the steps of image obtaining, image uploading, image recognizing and similarity verification, wherein in the image obtaining step, a target image is obtained, all ORB feature points of the target image are extracted through the ORB algorithm to generate an ORB description subsequence of the target image; in the image uploading step, the ORB description subsequence is uploaded to a cloud image database based on description subsamples; in the image recognizing step, the cloud image database conducts matching recognition on the image through the retrieval algorithm based on a vocabulary tree, and N candidate images with higher matching degree are fed back; in the similarity verification step, the candidate images are found in the cloud image database to obtain 128 dimensional vectors of the target image and the candidate images, the distances between the target image and the candidate images are calculated respectively, and the candidate image with the shortest distance is found out. By means of the method, influences of a poor network on recognition speed are small, the retrieval speed is high, and the retrieval precision is high.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a cloud image recognition method based on vocabulary tree retrieval and similarity verification. Background technique [0002] Real-time image search is a real-time image recognition technology that can support user-defined, ultra-large-scale image databases. It can realize real-time recognition of the image input content of the mobile terminal device. The whole recognition process is carried out in the cloud, so that users do not need to download the huge image database locally, and can also make full use of cloud computing resources for high-speed retrieval of the database. [0003] However, the current cloud-based real-time image recognition technology uploads local pictures to the server, and the server compares the pictures with recognition with the stored pictures one by one, which has the following defects: in the case of poor wireless networks, users upload ima...

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
IPC IPC(8): G06F17/30
CPCG06F16/51G06F16/583
Inventor 施茂燊
Owner 成都弥知科技有限公司
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