Target detection method and device

A detection method and a technology to be detected, which are applied in image data processing, instruments, biological neural network models, etc., and can solve problems such as difficulty in adapting to video target detection

Active Publication Date: 2019-06-11
TENCENT TECH (SHENZHEN) CO LTD
View PDF5 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Since the target detection method based on a single image in the prior art is difficult to adapt to the problem of target detection in vi

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
  • Target detection method and device
  • Target detection method and device
  • Target detection method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach

[0070] In a possible implementation manner, the product of the first mapping feature and the second mapping feature is determined, and then the index of the product is determined as the correlation matrix between the video image to be detected and the reference video image, specifically the following formula (1):

[0071]

[0072] Among them, f(x i , x j ) is the correlation matrix, x i is the first image feature of the video image to be detected, x j is the reference image feature of the reference video image, θ(x i ) is the first mapping feature, is the second mapping feature.

[0073] In a possible implementation manner, the product of the first mapping feature and the second mapping feature is determined as the correlation matrix between the video image to be detected and the reference video image, which may specifically be the following formula (2):

[0074]

[0075] Among them, f(x i , x j ) is the correlation matrix, xi is the first image feature of the v...

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 embodiment of the invention provides a target detection method and device. Relating to the technical field of machine learning, the method comprises the following steps of: obtaining a sample; themethod comprises the steps of detecting a target object in a video; extracting a first image feature from the to-be-detected video image by adopting a first target convolutional layer in the neural network model, wherein the to-be-detected video image is a frame of video image in the to-be-detected video, then determining association information between the first image feature and the reference image feature of the reference video image by adopting a non-local module in the neural network model, and then determining a target object in the to-be-detected video image based on the first image feature and the association information. Supplementing to the first image feature in the to-be-detected video image is performed by using the association information between the first image feature andthe reference image feature, therefore, more information related to the target object in the to-be-detected video image is obtained, the situations of object shielding and motion blur are effectivelyrelieved, and the precision of detecting the target object in the video is improved.

Description

technical field [0001] The embodiments of the present application relate to the technical field of machine learning, and in particular, to a target detection method and device. Background technique [0002] Image object detection refers to the position detection and classification of objects in the image. In target detection, a method of extracting target candidate boxes from the image to be detected or the feature map of the image to be detected is usually used, and the target candidate box is used to indicate a rectangular area where a target may exist. [0003] However, when performing target detection on a video, sometimes occlusions will occur between objects in a certain frame of image in the video, and sometimes objects moving at high speed in the video will appear motion blurred. At this time, target detection based on a single image methods are difficult to adapt to object detection in videos. Contents of the invention [0004] Since the target detection method ...

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): G06T7/73G06N3/04
Inventor 赵世杰李峰易阳邱日明
Owner TENCENT TECH (SHENZHEN) CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products