Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Portrait detection and segmentation method based on deep network context lifting

A deep network and context technology, applied in the field of image processing, can solve the problems of high cost of edge labeling, fuzzy fineness near the edge, etc., and achieve the effect of reducing the cost of data labeling

Active Publication Date: 2021-11-12
ZHUHAI FUDAN INNOVATION INST
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But in fact, the cost of edge labeling is expensive, and the edge calibration of most current data sets is based on manual portrait calibration, and the fineness near the edge is very vague.

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
  • Portrait detection and segmentation method based on deep network context lifting
  • Portrait detection and segmentation method based on deep network context lifting
  • Portrait detection and segmentation method based on deep network context lifting

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0055] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided for more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0056] seefigure 1 As shown, the embodiment of the present invention provides a method for portrait detection and segmentation based on deep network context promotion, which specifically includes the following steps:

[0057] S1. Based on the deep network framework, L depth features of different scales are extracted from the portrait image;

[0058] S2. Based on the highest scale feature, the depth feature of the highest scale is f...

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 portrait detection and segmentation method based on deep network context lifting, and the method specifically comprises the steps: extracting L depth features of different scales from a portrait image based on a deep network framework; on the basis of the highest-scale features, carrying out feature fusion on the highest-scale depth features on multiple pyramid scales through a pyramid pooling module, and generating global prior information; lifting and fusing context information of the depth features from a high scale to a low scale through a fusion block to obtain output features of each scale; optimizing and training the output features of each scale, and completing portrait detection and segmentation; through the method, deep network context information can be deeply mined from multiple scales, multiple spaces and multiple channels without the help of extra knowledge, and accurate portrait detection and segmentation of a monocular image are realized.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a method for detecting and segmenting portraits based on deep network context enhancement. Background technique [0002] Portrait detection and segmentation, as a special task of semantic segmentation, has a wide range of applications. For beautification applications, portrait detection is the basis for applications such as portrait image stylization, depth of field blur processing, and image matting; for security protection applications, portrait detection can blur processing and replace the background information of portrait images; portrait detection of monocular images is used in practical applications It is more important in the middle, because compared with the multi-camera image taken by the dual camera, it is not limited by the shooting light and distance. [0003] The main challenge of portrait detection based on deep learning is not only to accurately locate t...

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): G06T7/11G06T7/00G06K9/62
CPCG06T7/11G06T7/0002G06T2207/20081G06T2207/30196G06T2207/20016G06F18/253Y02D10/00
Inventor 许赢月王俊宇高自立
Owner ZHUHAI FUDAN INNOVATION INST
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com