We develop and analyze a new algorithm to find the connected components of a compact set I from a Lie group G endowed with a left-invariant Riemannian distance. For a given <InlineEquation ID="IEq1"> <EquationSource Format="TEX">$\delta >0$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>δ</mi> <mo>></mo> <mn>0</mn> </mrow> </math></EquationSource> </InlineEquation>, the algorithm finds the largest cover of I such that all sets in the cover are separated by at least distance <InlineEquation ID="IEq2"> <EquationSource Format="TEX">$\delta $</EquationSource> <EquationSource Format="MATHML"><math> <mi>δ</mi> </math></EquationSource> </InlineEquation>. We call the sets in the cover the <InlineEquation ID="IEq3"> <EquationSource Format="TEX">$\delta $</EquationSource> <EquationSource Format="MATHML"><math> <mi>δ</mi> </math></EquationSource> </InlineEquation>-connected components of I (closely related to <InlineEquation ID="IEq4"> <EquationSource Format="TEX">$\check{\text {C}}$</EquationSource> <EquationSource Format="MATHML"><math> <mover accent="true"> <mtext>C</mtext> <mo stretchy="false">ˇ</mo> </mover> </math></EquationSource> </InlineEquation>ech complexes of radius <InlineEquation ID="IEq5"> <EquationSource Format="TEX">$\delta /2$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>δ</mi> <mo stretchy="false">/</mo> <mn>2</mn> </mrow> </math></EquationSource> </InlineEquation>). The grouping relies on an iterative procedure involving morphological dilations with Hamilton-Jacobi-Bellman kernels on G and notions of <InlineEquation ID="IEq6"> <EquationSource Format="TEX">$\delta $</EquationSource> <EquationSource Format="MATHML"><math> <mi>δ</mi> </math></EquationSource> </InlineEquation>-thickened sets. We prove that the algorithm converges in finitely many iteration steps. We find the optimal value for <InlineEquation ID="IEq7"> <EquationSource Format="TEX">$\delta $</EquationSource> <EquationSource Format="MATHML"><math> <mi>δ</mi> </math></EquationSource> </InlineEquation> using persistence diagrams. We also propose to use specific affinity matrices. This allows for grouping of <InlineEquation ID="IEq8"> <EquationSource Format="TEX">$\delta $</EquationSource> <EquationSource Format="MATHML"><math> <mi>δ</mi> </math></EquationSource> </InlineEquation>-connected components based on their local proximity and alignment. Among the many different applications of the algorithm, in this article, we focus on illustrating that the method can efficiently identify (possibly overlapping) branches in complex vascular trees on retinal images. This is done by applying an orientation score transform to the images that allows us to view them as functions from <InlineEquation ID="IEq9"> <EquationSource Format="TEX">$\mathbb {L}_2(G)$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msub> <mi mathvariant="double-struck">L</mi> <mn>2</mn> </msub> <mrow> <mo stretchy="false">(</mo> <mi>G</mi> <mo stretchy="false">)</mo> </mrow> </mrow> </math></EquationSource> </InlineEquation> where <InlineEquation ID="IEq10"> <EquationSource Format="TEX">$G=SE(2)$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mi>G</mi> <mo>=</mo> <mi>S</mi> <mi>E</mi> <mo stretchy="false">(</mo> <mn>2</mn> <mo stretchy="false">)</mo> </mrow> </math></EquationSource> </InlineEquation>, the Lie group of roto-translations. By applying our algorithm in this Lie group, we illustrate that we obtain <InlineEquation ID="IEq11"> <EquationSource Format="TEX">$\delta $</EquationSource> <EquationSource Format="MATHML"><math> <mi>δ</mi> </math></EquationSource> </InlineEquation>-connected components that differentiate between crossing structures and that group well-aligned, nearby structures. This contrasts standard connected component algorithms in <InlineEquation ID="IEq12"> <EquationSource Format="TEX">$\mathbb {R}^2$</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mrow> <mi mathvariant="double-struck">R</mi> </mrow> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation>.

Connected Components on Lie Groups and Applications to Multi-Orientation Image Analysis

期刊论文
出版模式：

开放获取
发表日期：

2026年3月26日

查看全文

Nicky J. van den Berg,
Olga Mula,
Leanne Vis,
Remco Duits

摘要

We develop and analyze a new algorithm to find the connected components of a compact set I from a Lie group G endowed with a left-invariant Riemannian distance. For a given $\delta >0$ $δ > 0$ , the algorithm finds the largest cover of I such that all sets in the cover are separated by at least distance $\delta $ $δ$ . We call the sets in the cover the $\delta $ $δ$ -connected components of I (closely related to $\check{\text {C}}$ $\overset{ˇ}{C}$ ech complexes of radius $\delta /2$ $δ / 2$ ). The grouping relies on an iterative procedure involving morphological dilations with Hamilton-Jacobi-Bellman kernels on G and notions of $\delta $ $δ$ -thickened sets. We prove that the algorithm converges in finitely many iteration steps. We find the optimal value for $\delta $ $δ$ using persistence diagrams. We also propose to use specific affinity matrices. This allows for grouping of $\delta $ $δ$ -connected components based on their local proximity and alignment. Among the many different applications of the algorithm, in this article, we focus on illustrating that the method can efficiently identify (possibly overlapping) branches in complex vascular trees on retinal images. This is done by applying an orientation score transform to the images that allows us to view them as functions from $\mathbb {L}_2(G)$ $L_{2} (G)$ where $G=SE(2)$ $G = S E (2)$ , the Lie group of roto-translations. By applying our algorithm in this Lie group, we illustrate that we obtain $\delta $ $δ$ -connected components that differentiate between crossing structures and that group well-aligned, nearby structures. This contrasts standard connected component algorithms in $\mathbb {R}^2$ $R^{2}$ .