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

A semantic segmentation method and system based on continuous learning

A technology of semantic segmentation and learning algorithm, applied in the field of semantic segmentation methods and systems based on continuous learning, can solve the problems of inability to obtain, forget, and consistent performance, and achieve the effect of improving average performance and alleviating the problem of catastrophic forgetting.

Active Publication Date: 2022-04-08
TSINGHUA UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

When new observations (i.e., upcoming scenes) are distributed differently from old observations (i.e., past scenes), lane line detection models tend to overfit new observations and forget to learn from old observations. results in knowledge learned and thus cannot achieve consistent performance across all complex real-world scenarios

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
  • A semantic segmentation method and system based on continuous learning
  • A semantic segmentation method and system based on continuous learning
  • A semantic segmentation method and system based on continuous learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0046] Before introducing the embodiments of the present invention, the relevant terms involved in the embodiments of the present invention are first explained as follows:

[0047] RGB image: refers to the RGB image collected by the vehicle camera, which is a three-channel image.

[0048] Scene: refers to the scene generated by the environment change during the driving process of the vehicle. For example, a crowded scene when there are many vehicles around, a flas...

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 semantic segmentation method and system based on continuous learning. The method includes: collecting RGB images of the road surface in real time through a vehicle-mounted monocular camera; identifying the type of the current scene based on the RGB image, and obtaining its corresponding The optimal skeleton model; input the RGB image collected in real time into the optimal skeleton model, and output the target detection result. The method of the present invention improves the average performance of the existing skeleton model of target detection under complex multi-scenes.

Description

technical field [0001] The invention belongs to the technical field of automatic driving, and in particular relates to a semantic segmentation method and system based on continuous learning. Background technique [0002] Perception of the vehicle's surrounding environment plays an important role in autonomous driving. As an important perception technology, lane line detection provides the vehicle with the accurate position of each lane during automatic driving to ensure the safety of passengers and pedestrians. [0003] The current lane detection work mainly has the following limitations: (1) The data imbalance problem in different scenarios makes it difficult for the algorithm to obtain consistent performance in all scenarios. The lane line detection model will overfit when the amount of data is the largest, and it will get poor results when the amount of data is small; (2) The number and type of lanes in the real scene change dynamically, This makes it necessary to use a...

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): G06V20/56G06V10/26G06V10/774G06K9/62G06N20/00
CPCG06N20/00G06F18/214
Inventor 张新钰李骏李志伟刘华平刘玉超韩威
Owner TSINGHUA UNIV
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