Image-based target point position detection method and apparatus, and electronic device

A target point and point position technology, which is applied in the field of image-based target point position detection methods, devices, electronic equipment and storage media, and can solve problems such as low target point position detection performance.

Active Publication Date: 2019-07-30
BEIJING SANKUAI ONLINE TECH CO LTD
View PDF13 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The embodiment of the present application provides an image-based target point position detection method, which helps to improve the problem of low target point position detection performance

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-based target point position detection method and apparatus, and electronic device
  • Image-based target point position detection method and apparatus, and electronic device
  • Image-based target point position detection method and apparatus, and electronic device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0027] An image-based target point position detection method disclosed in this embodiment, such as figure 1 As shown, the method includes: Step 110 to Step 130.

[0028] Step 110, acquiring several image samples marked with the position of the real target point.

[0029] Before detecting the position of the target point through the neural network model, it is first necessary to construct a training sample to train the neural network model. In the embodiment of the present application, the detection of the key points of the face is taken as an example to illustrate the specific implementation of the method for detecting the position of the target point based on the image. Therefore, it is necessary to construct the training samples of the neural network model based on the face image. That is, the step of obtaining several image samples marked with real target point positions includes: obtaining several image samples marked with real face key point positions.

[0030] In some ...

Embodiment 2

[0059] An image-based target point position detection device disclosed in the embodiment of the present application, such as Figure 5 As shown, the device includes:

[0060] An image sample acquisition module 510, configured to acquire a number of image samples marked with real target point positions;

[0061] The model training module 520 is used to calculate the loss value of the neural network model according to variables through the loss function of the neural network model, and adjust the parameters of the neural network model with the goal of minimizing the loss value to train the neural network model. Network model; wherein, the loss function is configured to increase monotonically relative to the absolute value of the variable, the gradient function corresponding to the loss function is a monotonically decreasing function and the value of the gradient function is greater than 1 and when the variable approaches When infinity gradually approaches 1; the variable is use...

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 an image-based target point position detection method, belongs to the technical field of computers, and is used for solving the problem of low target point position detection performance. The point position detection method disclosed by the embodiment of the invention comprises the following steps: acquiring a plurality of image samples marked with real target point positions; calculating a loss value of the neural network model according to a variable through a loss function of the neural network model, adjusting parameters of the neural network model by taking the minimum loss value as a target, and training the neural network model; wherein the loss function is configured to monotonically increase relative to the absolute value of a variable, the gradient function corresponding to the loss function is a monotonically decreasing function, the value of the gradient function is greater than 1, and the gradient function gradually approaches 1 when the variable approaches infinite; wherein the variable is used for representing a difference value between a true value and a predicted value, the true value is determined based on a sample label, and the predictedvalue is predicted through the neural network model.

Description

technical field [0001] The present application relates to the technical field of image recognition and detection, in particular to an image-based target point position detection method, device, electronic equipment and storage medium. Background technique [0002] Image-based target point position detection has a wide range of applications, such as face key point detection based on face images, and pedestrian detection based on video images. Taking face detection as an example, face detection is the core technology of products such as face recognition, face beauty makeup, and face mapping. Many applications can be developed based on the key points of the face detected in the face image. With the continuous deepening of people's research on deep learning, the research on image-based target point position detection through deep learning methods has been continuously carried out, and certain results have been achieved. The usual target point position detection task is a regre...

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 Applications(China)
IPC IPC(8): G06K9/00G06K9/62
CPCG06V40/161G06F18/214
Inventor 柴振华赖申其
Owner BEIJING SANKUAI ONLINE TECH 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