Topology-Preserving Multi-label Image Segmentation

Waggoner, Jarrell; Zhou, Yunxuan; Simmons, Jeff; Graef, Marc De; Wang, Song

doi:10.1109/wacv.2015.149

Cited by 7 publications

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Topological information has been used in various learning and vision tasks. Examples include but are not limited to shape analysis (Reininghaus et al 2015;Carriere, Cuturi, and Oudot 2017), graph learning (Hofer et al 2017;Zhao and Wang 2019;Zhao et al 2020), clustering (Ni et al 2017;Chazal et al 2013), learning with label noise (Wu et al 2020) and image segmentation (Wu et al 2017;Mosinska et al 2018;Chan et al 2017;Waggoner et al 2015). Persistent-homology-based objective functions have been used for image segmentation (Hu et al 2019;Clough et al 2019), generative adversarial networks (GANs) (Wang et al 2020b), graphics (Poulenard, Skraba, and Ovsjanikov 2018) and machine learning model regularization (Hofer et al 2019;Chen et al 2019).…”

Section: Related Workmentioning

confidence: 99%

Localization in the Crowd with Topological Constraints

Abousamra

Hoai

Samaras

et al. 2021

AAAI

View full text Add to dashboard Cite

We address the problem of crowd localization, i.e., the prediction of dots corresponding to people in a crowded scene. Due to various challenges, a localization method is prone to spatial semantic errors, i.e., predicting multiple dots within a same person or collapsing multiple dots in a cluttered region. We propose a topological approach targeting these semantic errors. We introduce a topological constraint that teaches the model to reason about the spatial arrangement of dots. To enforce this constraint, we define a persistence loss based on the theory of persistent homology. The loss compares the topographic landscape of the likelihood map and the topology of the ground truth. Topological reasoning improves the quality of the localization algorithm especially near cluttered regions. On multiple public benchmarks, our method outperforms previous localization methods. Additionally, we demonstrate the potential of our method in improving the performance in the crowd counting task.

show abstract

Section: Related Workmentioning

confidence: 99%

Localization in the Crowd with Topological Constraints

Abousamra

Hoai

Samaras

et al. 2021

AAAI

View full text Add to dashboard Cite

show abstract

“…However, none of these methods o ers explicit control over the topology. While topological errors can be avoided by tting or evolving a template shape with the correct topology [Waggoner et al 2015], these methods are limited to the availability of templates.…”

Section: Modeling Multi-labeled Domainsmentioning

confidence: 99%

Topology-controlled reconstruction of multi-labelled domains from cross-sections

Huang¹,

Zou²,

Carr

et al. 2017

ACM Trans. Graph.

View full text Add to dashboard Cite

show abstract

“…Another example is the segmentation of objects with complicated interiors, noises, or occlusions, where a topological constraint can be used to prevent over-segmentation, i.e., the forming of "holes" due to image complexity [2], or under-segmentation, i.e., when the contours of separate objects merge. Much active research is undergone in the area, such as image segmentation and registration using the Beltrami representation of shapes [3] and non-local shape descriptors [4,5], multi-label image segmentation with preserved topology [6], and min-cut/max-flow segmentation using topology priors [7].…”

Section: Introductionmentioning

confidence: 99%

Using the Split Bregman Algorithm to Solve the Self-repelling Snakes Model

et al. 2022

View full text Add to dashboard Cite

Preserving contour topology during image segmentation is useful in many practical scenarios. By keeping the contours isomorphic, it is possible to prevent over-segmentation and under-segmentation, as well as to adhere to given topologies. The Self-repelling Snakes model (SR) is a variational model that preserves contour topology by combining a non-local repulsion term with the geodesic active contour model. The SR is traditionally solved using the additive operator splitting (AOS) scheme. In our paper, we propose an alternative solution to the SR using the Split Bregman method. Our algorithm breaks the problem down into simpler sub-problems to use lower-order evolution equations and a simple projection scheme rather than re-initialization. The sub-problems can be solved via fast Fourier transform or an approximate soft thresholding formula which maintains stability, shortening the convergence time, and reduces the memory requirement. The Split Bregman and AOS algorithms are compared theoretically and experimentally.

show abstract

Topology-Preserving Multi-label Image Segmentation

Cited by 7 publications

References 13 publications

Localization in the Crowd with Topological Constraints

Localization in the Crowd with Topological Constraints

Topology-controlled reconstruction of multi-labelled domains from cross-sections

Using the Split Bregman Algorithm to Solve the Self-repelling Snakes Model

Contact Info

Product

Resources

About