Title:
ACTIVE CONTOUR- AND DEEP LEARNING-BASED AUTOMATIC SEGMENTATION METHOD FOR FUZZY BOUNDARY IMAGE
Document Type and Number:
WIPO Patent Application WO/2021/047684
Kind Code:
A1
Abstract:
Disclosed in the present invention is an active contour- and deep learning-based automatic segmentation method for a fuzzy boundary image. In the method, first, a deep convolutional neural network model is used to segment a fuzzy boundary image to obtain an initial segmentation result; then, a contour in an inner region of the image segmented by the deep convolutional neural network model is used as an initialized contour and a contour constraint for an active contour model; and the active contour model, by means of an image feature of a surrounding region of each contour point, a contour to move towards a target boundary, and a precise segmentation line is obtained between a target region and other background regions. In the present invention, an active contour model is added to a base of a deep convolutional neural network model to further refine a segmentation result for a fuzzy boundary image, thus having the ability to segment a fuzzy boundary in an image, and further improving segmentation accuracy in a fuzzy boundary image.
More Like This:
Inventors:
CHEN JUNYING (CN)
YOU HAIJUN (CN)
YOU HAIJUN (CN)
Application Number:
PCT/CN2020/125703
Publication Date:
March 18, 2021
Filing Date:
October 31, 2020
Export Citation:
Assignee:
UNIV SOUTH CHINA TECH (CN)
International Classes:
G06T7/12; G06T7/11
Foreign References:
| CN110689545A | 2020-01-14 | |||
| CN108013904A | 2018-05-11 | |||
| CN106447688A | 2017-02-22 | |||
| CN103886564A | 2014-06-25 | |||
| CN106056576A | 2016-10-26 | |||
| US20180330477A1 | 2018-11-15 |
Attorney, Agent or Firm:
YOGO PATENT & TRADEMARK AGENCY LIMITED COMPANY (CN)
Download PDF: