On a variational and convex model of the Blake–Zisserman type for segmentation of low-contrast and piecewise smooth images

Burrows, Liam and Theljani, Anis and Chen, Ke (2021) On a variational and convex model of the Blake–Zisserman type for segmentation of low-contrast and piecewise smooth images. Journal of Imaging, 7 (11). 228. ISSN 2313-433X (https://doi.org/10.3390/jimaging7110228)

[thumbnail of Burrows-etal-JI-2021-On-a-variational-and-convex-model-of-the-Blake-Zisserman-type-for-segmentation-of-low-contrast]
Preview
 Text. Filename: Burrows-etal-JI-2021-On-a-variational-and-convex-model-of-the-Blake-Zisserman-type-for-segmentation-of-low-contrast.pdf
Final Published Version
License: Creative Commons Attribution 4.0 logo
Download (16MB)| Preview

Abstract

This paper proposes a new variational model for segmentation of low-contrast and piecewise smooth images. The model is motivated by the two-stage image segmentation work of Cai– Chan–Zeng (2013) for the Mumford–Shah model. To deal with low-contrast images more effectively, especially in treating higher-order discontinuities, we follow the idea of the Blake–Zisserman model instead of the Mumford–Shah. Two practical ideas are introduced here: first, a convex relaxation idea is used to derive an implementable formulation, and second, a game reformulation is proposed to reduce the strong dependence of coupling parameters. The proposed model is then analysed for existence and further solved by an ADMM solver. Numerical experiments can show that the new model outperforms the current state-of-the-art models for some challenging and low-contrast images.

  • Item type:Article
    ID code:87181
    Dates:
    Date
    Event
    28 October 2021
    Published
    22 October 2021
    Accepted
    6 September 2021
    Submitted
    Subjects:Science > Mathematics > Electronic computers. Computer science
    Department:UNSPECIFIED
    Depositing user: Pure Administrator
    Date deposited:06 Nov 2023 14:43
    Last modified:09 Apr 2024 04:22
    URI:https://strathprints.strath.ac.uk/id/eprint/87181
CORE (COnnecting REpositories)