Saved in:
Bibliographic Details
Main Authors: Scherzer, Otmar, Shi, Cong, Vu, Thi Lan Nhi
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.16516
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917486066139136
author Scherzer, Otmar
Shi, Cong
Vu, Thi Lan Nhi
author_facet Scherzer, Otmar
Shi, Cong
Vu, Thi Lan Nhi
contents Chan-Vese algorithms have proven to be a first-class method for image segmentation. Early implementations used level set methods with a pixelwise representation of the level set function. Later, parametrized level set approximations, such as splines, have been studied and computationally developed to improve efficiency. In this paper, we use neural networks as parametrized approximations of level set functions for implementing the Chan-Vese methods. We show that this approach is efficient because of the equivalence between two layer neural networks and polygonal approximations of level set-based segmentations. In turn, this allows the two-layer network architecture to be interpreted as an ansatz function for the approximate minimization of Chan-Vese functionals. Based on these theory, we extend the classical Chan-Vese algorithm to a data-driven setting, where prior parameters of the network are obtained through unsupervised training on representative image data. These learned parameters encode geometric structures of the data, leading to improved initialization and faster convergence of the Chan-Vese image segmentation.
format Preprint
id arxiv_https___arxiv_org_abs_2603_16516
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Neural network parametrized level sets for image segmentation
Scherzer, Otmar
Shi, Cong
Vu, Thi Lan Nhi
Numerical Analysis
Chan-Vese algorithms have proven to be a first-class method for image segmentation. Early implementations used level set methods with a pixelwise representation of the level set function. Later, parametrized level set approximations, such as splines, have been studied and computationally developed to improve efficiency. In this paper, we use neural networks as parametrized approximations of level set functions for implementing the Chan-Vese methods. We show that this approach is efficient because of the equivalence between two layer neural networks and polygonal approximations of level set-based segmentations. In turn, this allows the two-layer network architecture to be interpreted as an ansatz function for the approximate minimization of Chan-Vese functionals. Based on these theory, we extend the classical Chan-Vese algorithm to a data-driven setting, where prior parameters of the network are obtained through unsupervised training on representative image data. These learned parameters encode geometric structures of the data, leading to improved initialization and faster convergence of the Chan-Vese image segmentation.
title Neural network parametrized level sets for image segmentation
topic Numerical Analysis
url https://arxiv.org/abs/2603.16516