Method for RGB-D image segmentation

Tukun Li; Zhijie Xu; Chaolong Zhang; Yanglong Cao; Qiuyan Gai; Xie Pan

Method for RGB-D image segmentation

Tukun Li (Inventor), Zhijie Xu (Inventor), Chaolong Zhang (Inventor), Yanglong Cao (Inventor), Qiuyan Gai (Inventor), Xie Pan (Inventor)

Research output: Patent

Abstract

The invention discloses a method for RGB-D image segmentation, which is based on the fact that the method carries out semantic level segmentation on an RGB-D color depth image, extracts foreground pixel information in the depth image and takes a weight matrix as output for weighting and assisting feature information extraction in the RGB image, so that important feature information under two modes mutually complements and responds, the loss of the feature information in the sampling process is fully compensated, the feature expression in the semantic or space fineness is improved, and particularly the method has stronger sensitivity on object edge information.

Original language	English
Patent number	CN114723951B
Priority date	8/06/22
Publication status	Published - 4 Nov 2022

Access to Document

https://patents.google.com/patent/CN114723951B/en?oq=CN114723951B

Cite this

@misc{a385262324d443ed965094afc8097f41,

title = "Method for RGB-D image segmentation",

abstract = "The invention discloses a method for RGB-D image segmentation, which is based on the fact that the method carries out semantic level segmentation on an RGB-D color depth image, extracts foreground pixel information in the depth image and takes a weight matrix as output for weighting and assisting feature information extraction in the RGB image, so that important feature information under two modes mutually complements and responds, the loss of the feature information in the sampling process is fully compensated, the feature expression in the semantic or space fineness is improved, and particularly the method has stronger sensitivity on object edge information.",

author = "Tukun Li and Zhijie Xu and Chaolong Zhang and Yanglong Cao and Qiuyan Gai and Xie Pan",

year = "2022",

month = nov,

day = "4",

language = "English",

type = "Patent",

note = "CN114723951B",

}

TY - PAT

T1 - Method for RGB-D image segmentation

AU - Li, Tukun

AU - Xu, Zhijie

AU - Zhang, Chaolong

AU - Cao, Yanglong

AU - Gai, Qiuyan

AU - Pan, Xie

PY - 2022/11/4

Y1 - 2022/11/4

N2 - The invention discloses a method for RGB-D image segmentation, which is based on the fact that the method carries out semantic level segmentation on an RGB-D color depth image, extracts foreground pixel information in the depth image and takes a weight matrix as output for weighting and assisting feature information extraction in the RGB image, so that important feature information under two modes mutually complements and responds, the loss of the feature information in the sampling process is fully compensated, the feature expression in the semantic or space fineness is improved, and particularly the method has stronger sensitivity on object edge information.

AB - The invention discloses a method for RGB-D image segmentation, which is based on the fact that the method carries out semantic level segmentation on an RGB-D color depth image, extracts foreground pixel information in the depth image and takes a weight matrix as output for weighting and assisting feature information extraction in the RGB image, so that important feature information under two modes mutually complements and responds, the loss of the feature information in the sampling process is fully compensated, the feature expression in the semantic or space fineness is improved, and particularly the method has stronger sensitivity on object edge information.

M3 - Patent

M1 - CN114723951B

ER -

Method for RGB-D image segmentation

Abstract

Access to Document

Fingerprint

Cite this