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

A semantic segmented convolutional neural network with context information coding

A convolutional neural network and semantic segmentation technology, applied in the field of semantic segmentation convolutional neural network, can solve the problem of limited expression ability of context features, and achieve the effect of enriching global feature expression, avoiding unclear segmentation edges, and improving segmentation effect.

Inactive Publication Date: 2018-12-28
TIANJIN UNIV
View PDF3 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method uses vectors as context encoding, resulting in limited expressiveness of context features

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 segmented convolutional neural network with context information coding
  • A semantic segmented convolutional neural network with context information coding
  • A semantic segmented convolutional neural network with context information coding

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0030] The technical solutions in the embodiments of the present invention will be clearly and completely described below, and the convolutional neural network used for image semantic segmentation will be used as an example in the description. Obviously, the described embodiments are only some examples of the present invention, and Not all instances.

[0031] This section will start with (Badrinarayanan, Vijay, Alex Kendall, and Roberto Cipolla.2017. "SegNet: A Deep Convolutional Encoder-Decoder Architecture for Scene Segmentation." IEEE Transactions on Pattern Analysis and Machine Intelligence 39(12):2481–95. https: / / doi.org / 10.1109 / TPAMI.2016.2644615.) proposed SegNet method as a basis, obviously, the present invention does not limit the infrastructure, which is only an example.

[0032] (1) Prepare suitable training data. The training data in this example includes training images and pixel-by-pixel semantic segmentation annotations.

[0033](2) Build the basic network, th...

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 relates to a semantic segmented convolutional neural network with context information coding, which reconstructs the existing SegNet semantic segmented convolutional neural network. Thesteps are as follows: a semantic segmented convolutional neural network structure being built based on SegNet; selecting a position where global features are extracted and a position where global features are fused; determining a unified dimension of the global eigentensor; determining the construction of each global feature extractor and each global feature fusion; the new global features extracted by each global feature extractor being added with the existing global features element by element. SegNet with global information encoding is obtained by connecting the global feature output from each global feature fuser with the current local feature as new feature information. A semantic segmented convolutional neural network with global information coding is obtained.

Description

technical field [0001] The invention belongs to the field of machine learning and neural network, in particular to a semantic segmentation convolutional neural network with context information coding. Background technique [0002] Convolutional Neural Network (CNN) has achieved the best results in semantic segmentation tasks of images and videos. The mainstream network structure includes full-volume and network-like structures and encoder-decoder structures. There have been some methods to model context information in semantic segmentation networks, but the effect is very limited. [0003] Criteria Random Field (CRF) is a post-processing method that can model contextual information (Shuai Zheng et al.2015; and Koltun 2011; Chen, Papandreou, Kokkinos, et al. 2017), but require an additional iterative step to train the CRF. In addition, as a post-processing method, CRF is not helpful for neural network feature extraction and expression. [0004] ParseNets (Liu, Rabinovich...

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): G06K9/34G06K9/46
CPCG06V10/267G06V10/44
Inventor 庞彦伟孙汉卿
Owner TIANJIN 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