Method for estimating shape parameters of the fovea by optical coherence tomography

1. A method for estimating shape parameters of an image data set ( 1 ) of a fovea ( 100 ), comprising the steps of:

Acquiring an image data set ( 1 ) of the macula ( 50 ) of an eye, the macula comprising a foveal pit ( 101 ) and a foveal rim ( 102 ),

Estimating the center of the foveal pit ( 101 ) from the image data set ( 1 ),

Estimating from the center of the foveal pit ( 101 ), radially extending height profiles (c) of the fovea ( 100 ),

For each height profile (c), fitting a model-function to the height profile (c),

Estimating for each height profile (c) a set of fit parameters from the fitted model-function, and

Determining from the sets of fit parameters at least one shape parameter of the fovea ( 100 ),

wherein the model function is a smooth composite Bézier path comprising cubic Bézier functions, and

wherein each height profile (c) comprises an interior segment (c I ) and an exterior segment (c E ), wherein the height profiles (c) are fitted with the smooth composite Bézier path, wherein a first Bézier function (Q I ) is fitted to the interior segment (c I ) and wherein a second Bézier function (Q E ) is fitted to the exterior segment (c E ), wherein the interior segment (c I ) extends between a first end point (P 0 I ) and a second end point (P 3 I ) of the first Bézier function (Q I ), and the exterior segment (c E ) extends between a first end point (P 0 E ) and a second end point (P 3 E ) of the second Bézier function (Q E ), wherein the second end point (P 3 I ) of the first Bézier function (Q I ) is identical or adjoining with the first end point (P 0 E ) of the second Bézier function (Q E ), wherein the first end point (P 0 I ) of the first Bézier function (Q I ) is at a first end of the height profile (c), the second end point (P 3 I ) of the first Bézier function (Q I ) and the first end point (P 0 E ) of the second Bézier function (Q E ) are identical at a maximum rim height of the height profile (c) and the second end point (P 3 E ) of the second Bézier function (B 2 ) is at a second end of the height profile (c), wherein the first and second Bézier function (Q I , Q E ) have the same slope at the second end point (P 3 I ) of the first Bézier function (Q I ) and the identical first end point (P 0 E ) of the second Bézier function (Q E ).

2. Method according to claim 1 , wherein the height profiles (c) each comprise segments, wherein each segment is fitted with the Bézier function comprised by the smooth composite Bézier path.

3. Method according to claim 1 , wherein each height profile (c) is determined from a surface data set ( 200 ), wherein the surface data set ( 200 ) comprises a distance between the inner limiting membrane ( 103 ) representing data and the retinal pigment epithelium layer ( 104 ) representing data comprised in the image data set ( 100 ), wherein the height profiles (c) are line profiles along the respective radial direction of the surface data set ( 200 ).

4. Method according to claim 1 , wherein the height profiles (c) are regularly spaced.

5. Method according to claim 1 , wherein the at least one determined shape parameter of the fovea ( 100 ) is one of the following:

A central foveal thickness (h cft ),

An average rim height (h r )

A rim disk area (A r ),

A minor axis and a minor axis length (λ 2 r ) of the rim disk area (A r ),

A major axis and a major axis length (λ 3 r ) of the rim disk area (A r ),

An average rim disk diameter (d r ),

An average pit depth (h p ),

An average maximum pit slope (s m ),

An average slope disk diameter (d s ),

A slope disk area (A s ),

A minor axis and a minor axis length (λ 2 s ) of the slope disk area (A s ),

A major axis and a major axis length (λ 3 s ) of the slope disk area (A s ),

A pit disk area (A f ),

A minor axis and a minor axis length (λ 2 f ) of the pit disk area (A f ),

A major axis and a major axis length (λ 3 f ) of the pit disk area (A f ),

An average pit disk diameter (d f )

A rim volume (V r ),

An inner rim volume (V IR ), and/or

A pit volume (V p ).

6. Method according to claim 5 , wherein a first set of shape parameters associated to a first morphological group of the fovea ( 100 ) comprises:

The minor axis length (λ 2 s ) of the slope disk area (A s ),

The slope disk area (A s ), and

The average slope disk diameter (d s ).

7. Method according to claim 5 , wherein a second set of shape parameters associated to a second morphological group of the fovea ( 100 ) comprises:

The average rim height (h r ),

The major axis length (λ 3 s ) of the slope disk area (A s ),

The pit disk area (A f ),

The average pit disk diameter (d f ),

The minor axis length (λ 2 f ) of the pit disk area (A f ),

The inner rim volume (V IR ).

8. Method according to claim 5 , wherein a third set of shape parameters associated to a third morphological group of the fovea ( 100 ) comprises:

The pit depth (h p ),

The major axis length (λ 3 f ) of the pit disk area (A f ), and

The average maximum pit slope (s m ).

9. Method according to claim 5 , wherein a fourth set of shape parameters associated to a fourth morphological group of the fovea ( 100 ) comprises:

The central foveal thickness (h cft ),

The average rim diameter (d r ),

The rim disk area (A r ),

The major axis length (λ 3 r ) of the rim disk area (A r ),

The minor axis length (λ 2 r ) of the rim disk area (A r ),

The rim volume (V r ),

The pit volume (V f ), and

The inner rim volume (V IR ).

10. Method according to claim 5 , wherein the image data set ( 1 ) is assigned to a morphological group, wherein the image data set ( 1 ) is assigned based on the value of at least one of the shape parameters by comparing the magnitude or deviation of the respective shape parameter to a predefined reference value for the shape parameter.

11. Method according to claim 5 , wherein a trained classifier assigns the image data set ( 1 ) to a morphological group, wherein the image data set ( 1 ) is assigned based on the value of at least one of the shape parameters by comparing the magnitude or deviation of the respective shape parameter to a predefined reference value for the shape parameter.