Detection algorithm of animal face key point

A detection algorithm and key point technology, applied in computing, computer components, neural learning methods, etc., can solve the problems of few labeled samples, lack of evaluation database, etc., and achieve the effect of small model, weakened dependence, and fast speed

Active Publication Date: 2018-02-16
CHENDU PINGUO TECH
View PDF6 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The reason is that compared with the key points of human faces, there are fewer labeled samples of key points of animal faces, and there is a lack of public evaluation databases, etc.

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
  • Detection algorithm of animal face key point

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0018] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below.

[0019] Step 101, setting network structure, described network structure comprises bbox regression network and key point localization network;

[0020] The bbox regression network in this embodiment is composed of several convolution modules and fully connected layers. The main goal is to return a bbox that is more suitable for key point positioning. The output values ​​​​are the coordinates of the upper left corner and lower right corner of the rectangular box; key point positioning The network consists of several convolutional modules and fully connected layers, and its main goal is to predict the coordinates of key points on the animal's face.

[0021] Step 102, performing cascade training on the network structure;

[0022] The specific method includes: performing the rotation of the obtained training sam...

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 detection algorithm of an animal face key point. The detection algorithm comprises: a network structure is set; the network is divided into two sections, and a first sectionis returned to the bbox to obtain a face frame enabling an animal facial image to be inputted accurately; and the second section is returned to an animal face key point and the animal facial image andan animal facial image intercepted again based on the bbox are inputted, so that context information of the animal face is considered and the key region of the animal face is focused and thus the precise key point detection result is obtained. Besides, a stage-divided training mode is employed during the network training process; at the first stage, augmented samples with the large rotation angleand translation amplitude are used for training; and at the second stage, augmented samples with the small rotation angle and translation amplitude are used for training. The detection algorithm hasthe following advantages: the model is small and the speed is fast; a few of manually calibrated samples are needed; an animal face key point is detected accurately; and the need of real-time beauty makeup for a pet from a user is met.

Description

technical field [0001] The invention relates to the technical field of digital image processing, in particular to a detection algorithm for key points of animal faces. Background technique [0002] In recent years, selfie beauty makeup has received more and more attention, and the demand for cute pet beauty makeup has also emerged. Just as human face beauty makeup relies on the precise positioning of facial key points, cute pet beauty makeup also has a strong dependence on animal facial key points. Therefore, it is very important to explore a fast and accurate key location algorithm for animal faces. In addition, the research on animal facial key point positioning point algorithm can also be used in animal facial expression recognition, pain recognition and other aspects. [0003] At present, there are relatively few algorithms for locating key points on animal faces in both academia and industry. The reason is that compared with the key points of human faces, there are f...

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): G06K9/00G06N3/08
CPCG06N3/08G06V40/165G06V40/171
Inventor 陈丹
Owner CHENDU PINGUO TECH
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