Unlock instant, AI-driven research and patent intelligence for your innovation.

Image re-recognition method and recognition device

A re-identification and image technology, applied in the field of image re-recognition methods and recognition devices, can solve the problems of inexhaustibility, inability to detect image attributes, inability to overcome diversity, etc.

Active Publication Date: 2021-08-20
HANGZHOU HUAWEI ENTERPRISE COMM TECH
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] It can be seen that the existing image re-recognition technology needs to manually design a variety of attribute features, but in fact there are many types and quantities of attributes, which cannot be exhausted manually.
The existing image re-identification technology cannot overcome the diversity and cannot accurately detect the attribute characteristics of the image

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
  • Image re-recognition method and recognition device
  • Image re-recognition method and recognition device
  • Image re-recognition method and recognition device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0138] An embodiment of the present invention provides an image re-recognition method, such as Figure 7 As shown, the method includes the following steps:

[0139] 401. Receive a to-be-searched image, and determine a target image according to the to-be-searched image; the target image includes only the to-be-searched target.

[0140] In a specific implementation, the background of the image to be searched is removed, and only the target to be searched is retained; an image containing only the target to be searched is used as the target image.

[0141] Similarly, the background of the image to be searched is removed through the "moving object detection" algorithm, and the image to be searched is used by Pic i said, Mask i Indicates Pic i The background removal situation, when Mask i (x,y)=1 means Pic i (x,y) is the foreground, Mask i (x,y)=0 means Pic i (x, y) is the background, and subsequent operations are only performed on the foreground pixels of the image to be sea...

Embodiment 2

[0162] An embodiment of the present invention provides an identification device, such as Figure 8 As shown, the identification device includes: a receiving unit 501 , a determining unit 502 , an acquiring unit 503 , and a combining unit 504 .

[0163] The receiving unit 501 is configured to receive an image to be searched.

[0164] The determining unit 502 is configured to determine a target image according to the image to be searched; the target image only includes the target to be searched.

[0165] The obtaining unit 503 is configured to obtain a global feature vector of the target image and a local feature vector of the target image.

[0166] A combining unit 504, configured to combine the global feature vector and the local feature vector to form a feature vector set of the target image.

[0167] The obtaining unit 503 is further configured to obtain a set of feature vectors of historical images in a pre-generated image library, the image library is generated according...

Embodiment 3

[0199] An embodiment of the present invention provides an identification device, including: a processor 601 and a memory 602 .

[0200] Wherein, the processor 601 may be a central processing unit (English: central processing unit, abbreviated as CPU). The memory 602 is used to store program codes, and transmit the program codes to the processor 601, and the processor 601 executes the following instructions according to the program codes.

[0201] The memory 602 may include a volatile memory (English: volatile memory), such as a random access memory (English: random-access memory, abbreviated as RAM); the memory 602 may also include a non-volatile memory (English: non-volatile memory ), such as read-only memory (English: read-only memory, abbreviation: ROM), flash memory (English: flash memory), hard disk (English: hard disk drive, abbreviation: HDD) or solid state drive (English: solid-state drive, abbreviation: SSD). The memory 602 may also include a combination of the abov...

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 implementation of the present invention provides an image re-recognition method and recognition device, which relate to the field of image display, and can overcome the diversity of attribute features and accurately detect the attribute features of images. The method includes: receiving an image to be searched, determining a target image according to the image to be searched; the target image only includes the target to be searched; obtaining a global feature vector of the target image and a local feature vector of the target image; The feature vector is combined with the local feature vector to form the feature vector set of the target image; the feature vector set of the historical image in the pre-generated image library is obtained, and the image library is based on the global feature vector of the historical image and The local feature vectors of the historical images are generated; the historical images corresponding to the feature vector sets in the historical images that match the feature vector sets of the target image are determined as the matching images of the target image.

Description

technical field [0001] The invention relates to the field of image display, in particular to an image re-recognition method and a recognition device. Background technique [0002] In order to create a safe urban environment, surveillance video is widely used, and with the rapid growth of video data, image re-identification technology has received widespread attention. Such as: search for pictures by picture, that is, take the picture as input and find the matching picture in the video database. [0003] Existing image re-recognition technologies mainly include: artificially designing a variety of attribute features, and designing a corresponding detector for each of the designed attribute features. For example, the attribute can be - watch, hat, etc. A detector is designed for the attribute feature of "watch", and a detector is designed for the attribute feature of "hat". Each detector is used in turn to detect the attribute feature of an image, and the attribute feature o...

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): G06F16/583G06K9/46G06K9/62G06T7/10G06T3/00
CPCG06F16/5838G06T3/00G06T2207/10016G06T2207/30232G06V10/40G06V10/751
Inventor 骆立俊白博唐振陈茂林
Owner HANGZHOU HUAWEI ENTERPRISE COMM TECH