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

Human body segmentation neural network training method and equipment

A neural network and training method technology, applied in the training field of human body segmentation neural network, can solve the problem of unsmooth edges of human body segmentation images, and achieve the effect of less obvious edge jaggedness, smooth edges and good display effect.

Active Publication Date: 2022-04-22
BEIJING BYTEDANCE NETWORK TECH CO LTD
View PDF11 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Embodiments of the present disclosure provide a training method and equipment for a human body segmentation neural network to overcome the problem that the edge of the human body segmentation image obtained by the deep learning neural network trained by the existing mask is not correct when segmenting the human body area in the image. smooth problem

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
  • Human body segmentation neural network training method and equipment
  • Human body segmentation neural network training method and equipment
  • Human body segmentation neural network training method and equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] In order to make the purpose, technical solutions and advantages of the embodiments of the present disclosure clearer, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments It is a part of the embodiments of the present disclosure, but not all of them. Based on the embodiments in the present disclosure, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present disclosure.

[0030] refer to figure 1 , figure 1 It is a schematic diagram of the system architecture of the training method of the human body segmentation neural network provided by the embodiment of the present invention. like figure 1 As shown, the system provided in this embodiment includes a terminal device 101 and a server 102 . Wherein, the te...

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

Embodiments of the present disclosure provide a training method and equipment for a human body segmentation neural network. The method includes: labeling the first training image, wherein the pixel value of the human body segmentation area is marked as 1, and the pixel value of the background area is marked as 0; After the pixel value of the first training image is multiplied by 255, the second training image with pixel value 0 and 255 is obtained, and the pixel value of the second training image is interpolated to obtain the third training image with pixel value 0 to 255 ; The pixel values ​​of the third training image are divided by 255 respectively to obtain a mask with a pixel value of 0 to 1, and the mask with a pixel value of 0 to 1 is truncated according to a preset value to obtain a pixel value of 0 and A mask of 1; the mask with the truncated pixel value of 0 and 1 is input as the result data to the preset neural network training, when the trained preset neural network in the embodiment of the present disclosure is used to segment the human body area in the image , the edge of the obtained human body segmentation image is smoother.

Description

technical field [0001] Embodiments of the present disclosure relate to the field of computer technology, and in particular to a training method and equipment for a human body segmentation neural network. Background technique [0002] In the process of small video production, in order to meet the individual needs of customers, video software usually provides image fusion function: that is, to segment the human body area in the image, and then delete the original background area and replace it with a background with special effects to enhance the video quality. Beautification effect. [0003] At present, the segmentation of human body images is mainly realized through deep learning neural networks. The processing process is mainly as follows: downsampling the target image to a size corresponding to the input size of the deep learning neural network, and then using the deep learning neural network to segment the human body in the image. The region is identified, the human body...

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): G06T7/11G06T7/194
CPCG06T7/11G06T7/194G06T2207/20081G06T2207/20084G06T2207/30196
Inventor 李华夏
Owner BEIJING BYTEDANCE NETWORK TECH CO LTD