# Geodesic Length Measurement in Medical Images: Effect of the Discretization by the Camera Chip and Quantitative Assessment of Error Reduction Methods

^{*}

^{†}

## Abstract

**:**

## 1. Introduction

## 2. Methods

#### 2.1. In-Silico Model

- Choose a function (randomly).
- Choose a window width s (in pixels) in the image within the limits of [65 350] (randomly) (the mathematical function will be projected into this window).
- The function will be rotated with a random angle $\alpha $.
- The windowed and rotated mathematical function will be projected onto the resolution grid and thereby discretized.

#### 2.2. Centerline Extraction

#### 2.2.1. Erosion

#### 2.2.2. Voronoi Diagram

#### 2.2.3. Centerline Post-Processing

#### 2.3. Length Measurement

#### 2.3.1. Ground Truth Length

#### 2.3.2. Discrete Length of Centerline

#### 2.3.3. Continuous Length by Polynomial Approximation

#### 2.3.4. Continuous Length by Bézier Curve

#### 2.4. Physical Length Measurement

- Carl Zeiss Meditec AG PENTERO
^{®}900 (Surgical microscope) - Silicone tubes RCT THOMAFLUID
^{®} - Rotational plate (Thorlabs PR01(/M))
- Blood analog (52.4 mL demineralized water, 41.5 mL Glycerin (99.5%) and 6.7 g protein powder)
- ICG (PULSION Medical Systems SE)

#### 2.5. Evaluation

- Continuous ground truth length
- Discrete ground truth length without any centerline reconstruction
- Discrete length with the centerline reconstruction by erosion
- Discrete length with the centerline reconstruction by Voronoi diagram
- Continuous length with the centerline reconstruction by erosion and interpolation by Bézier
- Continuous length with the centerline reconstruction by erosion and interpolation by polynomial approximation
- Continuous length with the centerline reconstruction by Voronoi diagram and interpolation by Bézier
- Continuous length with the centerline reconstruction by Voronoi diagram and interpolation by polynomial approximation

## 3. Results

#### 3.1. In Silico Results

#### 3.2. Experimental Results

## 4. Discussion and Conclusions

## 5. Outlook

## Author Contributions

## Funding

## Acknowledgments

## Conflicts of Interest

## Abbreviations

ICG | Indocyanine Green |

ROI | Region of interest |

PAL | Phase Alternating Line |

## References

- Mendez, A. Optics in Medicine; Springer: Berlin/Heidelberg, Germany, 2016; pp. 299–337. [Google Scholar] [CrossRef][Green Version]
- Van Soest, G.; Regar, E.; van der Steen, A.F.W. Photonics in cardiovascular medicine. Nat. Photonics
**2015**, 9, 626–629. [Google Scholar] [CrossRef] - Baskar, R.; Lee, K.A.; Yeo, R.; Yeoh, K.W. Cancer and radiation therapy: Current advances and future directions. Int. J. Med. Sci.
**2012**, 9, 193–199. [Google Scholar] [CrossRef] [PubMed][Green Version] - Cubeddu, R.; Bassi, A.; Comelli, D.; Cova, S.; Farina, A.; Ghioni, M.; Rech, I.; Pifferi, A.; Spinelli, L.; Taroni, P.; et al. Photonics for life. IEEE Pulse
**2011**, 2, 16–23. [Google Scholar] [CrossRef] [PubMed] - Stummer, W.; Pichlmeier, U.; Meinel, T.; Wiestler, O.D.; Zanella, F.; Reulen, H.J. Fluorescence-guided surgery with 5-aminolevulinic acid for resection of malignant glioma: A randomised controlled multicentre phase III trial. Lancet Oncol.
**2006**, 7, 392–401. [Google Scholar] [CrossRef] - Zaunseder, S.; Trumpp, A.; Wedekind, D.; Malberg, H. Cardiovascular assessment by imaging photoplethysmography—A review. Biomed. Tech. Biomed. Eng.
**2018**, 63, 617–634. [Google Scholar] [CrossRef] [PubMed] - Raabe, A.; Beck, J.; Gerlach, R.; Zimmermann, M.; Seifert, V. Near-infrared indocyanine green video angiography: A new method for intraoperative assessment of vascular flow. Neurosurgery
**2003**, 52, 132–139. [Google Scholar] - Zygmont, M.E.; Lam, D.L.; Nowitzki, K.M.; Burton, K.R.; Lenchik, L.; McArthur, T.A.; Sekhar, A.K.; Itri, J.N. Opportunities for Patient-centered Outcomes Research in Radiology. Acad. Radiol.
**2016**, 23, 8–17. [Google Scholar] [CrossRef] - Molyneux, A.J.; Kerr, R.S.C.; Birks, J.; Ramzi, N.; Yarnold, J.; Sneade, M.; Rischmiller, J. Risk of recurrent subarachnoid haemorrhage, death, or dependence and standardised mortality ratios after clipping or coiling of an intracranial aneurysm in the International Subarachnoid Aneurysm Trial (ISAT): Long-term follow-up. Lancet Neurol.
**2009**, 8, 427–433. [Google Scholar] [CrossRef][Green Version] - Fantini, S.; Sassaroli, A.; Tgavalekos, K.T.; Kornbluth, J. Cerebral blood flow and autoregulation: Current measurement techniques and prospects for noninvasive optical methods. Neurophotonics
**2016**, 3, 031411. [Google Scholar] [CrossRef] - Amin-Hanjani, S.; Meglio, G.; Gatto, R.; Bauer, A.; Charbel, F.T. The utility of intraoperative blood flow measurement during aneurysm surgery using an ultrasonic perivascular flow probe. Neurosurgery
**2006**, 58, ONS-305–ONS-312. [Google Scholar] [CrossRef] - Transonic. Transonic Clinical Flowprobes Manual; Transonic Systems Inc.: Ithaca, NY, USA, 2019. [Google Scholar]
- Sugiyama, T. Basic Technology and Clinical Applications of the Updated Model of Laser Speckle Flowgraphy to Ocular Diseases. Photonics
**2014**, 1, 220–234. [Google Scholar] [CrossRef][Green Version] - Alander, J.T.; Kaartinen, I.; Laakso, A.; Pätilä, T.; Spillmann, T.; Tuchin, V.V.; Venermo, M.; Välisuo, P. A review of indocyanine green fluorescent imaging in surgery. Int. J. Biomed. Imaging
**2012**. [Google Scholar] [CrossRef] [PubMed] - Saito, M.; Saga, T.; Hayashi, H.; Noro, S.; Wada, H.; Kamada, K. Quantitative Blood Flow Assessment by Multiparameter Analysis of Indocyanine Green Video Angiography. World Neurosurg.
**2018**, 116, e187–e193. [Google Scholar] [CrossRef] [PubMed][Green Version] - Weichelt, C.; Duscha, P.; Steinmeier, R.; Meyer, T.; Kuß, J.; Cimalla, P.; Kirsch, M.; Sobottka, S.B.; Koch, E.; Schackert, G.; et al. Quantitative fluorescence angiography for neurosurgical interventions. Biomed. Tech. Biomed. Eng.
**2013**, 58, 269–279. [Google Scholar] [CrossRef] [PubMed] - Cimalla, P.; Graf, D.; Duscha, P.; Meyer, T.; Kuß, J.; Steinmeier, R.; Koch, E.; Morgenstern, U. Methods for determining the blood flow velocity in cerebral vessels using intraoperative Indocyanine Green fluorescence video angiography. In Proceedings of the 4th European Conference of the International Federation for Medical and Biological Engineering (IFMBE), Antwerp, Belgium, 23–27 November 2008; Volume 22, pp. 1446–1449. [Google Scholar] [CrossRef]
- Rieger, S.; Klee, S.; Baumgarten, D. Experimental Characterization and Correlation of Mayer Waves in Retinal Vessel Diameter and Arterial Blood Pressure. Front. Physiol.
**2018**, 9, 892. [Google Scholar] [CrossRef] [PubMed][Green Version] - Johnson Chacko, L.; Schmidbauer, D.T.; Handschuh, S.; Reka, A.; Fritscher, K.D.; Raudaschl, P.; Saba, R.; Handler, M.; Schier, P.P.; Baumgarten, D.; et al. Analysis of Vestibular Labyrinthine Geometry and Variation in the Human Temporal Bone. Front. Neurosci.
**2018**, 12, 107. [Google Scholar] [CrossRef][Green Version] - Cuesta Cuesta, A.B.; Martin Rios, M.D.; Noguero Meseguer, M.R.; Garcia Velasco, J.A.; de Matias Martinez, M.; Bartolome Sotillos, S.; Abreu Griego, E. Accuracy of tumor size measurements performed by magnetic resonance, ultrasound and mammography, and their correlation with pathological size in primary breast cancer. Cir. Esp.
**2019**, 97, 391–396. [Google Scholar] [CrossRef] - Long, F.; Ozturk, M.; Wolff, M.; Intes, X.; Kotha, S. Dental Imaging Using Mesoscopic Fluorescence Molecular Tomography: An ex Vivo Feasibility Study. Photonics
**2014**, 1, 488–502. [Google Scholar] [CrossRef] - Takahashi, Y.; Tsuchitani, T.; Kotoura, N.; Kitajima, K. Influence of image noise and object size on segmentation accuracy of FDG-PET imaging: A phantom experiment. Radiol. Phys. Technol.
**2019**, 12, 343–350. [Google Scholar] [CrossRef] - Bozzato, A.; Struffert, T.; Hertel, V.; Iro, H.; Hornung, J. Analysis of the accuracy of high-resolution computed tomography techniques for the measurement of stapes prostheses. Eur. Radiol.
**2010**, 20, 566–571. [Google Scholar] [CrossRef] - Griebel, J.; Dierking, W. Impact of Sea Ice Drift Retrieval Errors, Discretization and Grid Type on Calculations of Ice Deformation. Remote Sens.
**2018**, 10, 393. [Google Scholar] [CrossRef][Green Version] - Naber, A.; Berwanger, D.; Nahm, W. In Silico Modelling of Blood Vessel Segmentations for Estimation of Discretization Error in Spatial Measurement and its Impact on Quantitative Fluorescence Angiography. In Proceedings of the 41st Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), Berlin, Germany, 23–27 July 2019; pp. 4787–4790. [Google Scholar] [CrossRef]
- Lee, T.C.; Kashyap, R.L. Building skeleton models via 3-D medial surface/axis thinning algorithms. Graph. Model. Image Process.
**1994**, 56, 462–478. [Google Scholar] [CrossRef] - Zeng, C.; Bird, S.; Luce, J.; Wang, J. A Natural-Rule-Based-Connection (NRBC) Method for River Network Extraction from High-Resolution Imagery. Remote Sens.
**2015**, 7, 14055–14078. [Google Scholar] [CrossRef][Green Version] - Antiga, L. Patient-Specific Modeling of Geometry and Blood Flow in Large Arteries. Ph.D. Thesis, Politecnico Di Milano, Milano, Italy, 2002. [Google Scholar]
- Moon, T.K. Mathematical Methods and Algorithms for Signal Processing; Prentice Hall: Browse distinctive, NJ, USA, 2000. [Google Scholar]
- Kaschke, M.; Donnerhacke, K.H.; Rill, M.S. Optical Devices in Ophthalmology and Optometry: Technology, Design Principles and Clinical Applications; Wiley: Hoboken, NJ, USA, 2014. [Google Scholar] [CrossRef]

**Figure 1.**An overview of the methods used in this paper. Section 2.1, Section 2.2 and Section 2.3 deal with the in silico method and Section 2.4 with the experimental method.

**Figure 2.**To the

**left**the Voronoi diagram of a segmentation mask, to the

**right**a zoom into a section of this diagram—blue are the Voronoi edges, red is the centerline extracted from the Voronoi edges.

**Figure 3.**To the

**left**, an incorrect measurement due to a step-like connection with a distance of 2, to the

**right**, the corrected centerline with a measured distance of $\sqrt{2}$.

**Figure 4.**An exemplary image of a silicone tube filled with Indocyanine Green (ICG). It is fixed to a rotational board and the red line indicates 0° of rotation.

**Figure 5.**In silico simulation of two vessel segmentations.

**Left**: Section of a parabola with a stenosis.

**Right**: Bifurcation including sections of two Gaussian bells and a sinusoidal.

**Figure 6.**Relative length error using the erosion method in dependency on the angle for straight silicone tubes. The angle is measured to the horizontal pixel grid structure.

**Table 1.**Mathematical function, their formula and variables [25].

Function Type | Detailed Type | Formula | Variables’ Limit |
---|---|---|---|

Straight lines | $f\left(x\right)=tan\left(\theta \right)\xb7x$ | $\theta =[-{45}^{\circ},{45}^{\circ}]$, $x\phantom{\rule{0.222222em}{0ex}}is\phantom{\rule{0.222222em}{0ex}}arbitrary$ | |

Parabolas | Left arm | $f\left(x\right)=a\phantom{\rule{3.33333pt}{0ex}}\xb7{x}^{2}$ | $a=[1,2]$ |

${x}_{low}=[-\sqrt{2s},-2]$, ${x}_{high}=[{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}2,0]$ | |||

Right arm | $a=[1,2]$ | ||

${x}_{low}=[0,\sqrt{2s}-2]$, ${x}_{high}=[{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}2,\sqrt{2s}]$ | |||

Vertex | $a=[1,2]$ | ||

${x}_{low}=[-\sqrt{2s},-1]$, ${x}_{high}=[1,\sqrt{2s}]$ | |||

L.a. inverted | $a=[-1,-2]$ | ||

${x}_{low}=[-\sqrt{2s},-3]$, ${x}_{high}=[{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}2,1]$ | |||

R.a. inverted | $a=[-1,-2]$ | ||

${x}_{low}=[1,\sqrt{2s}-2]$, ${x}_{high}=[{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}2,\sqrt{2s}]$ | |||

V. inverted | $a=[-1,-2]$ | ||

${x}_{low}=[-\sqrt{2s},-2]$, ${x}_{high}=[2,\sqrt{2s}]$ | |||

Polynomials | $f\left(x\right)={\sum}_{k=4}^{n}{a}_{k}\xb7{x}^{k}$ | $n\le 15$, $x=[-5,5]$ | |

Sinusoids | Low frequency | $f\left(x\right)=a\xb7sin(f\xb7\pi \xb7x)$ | $a=[0.2s,0.7s]$, $f=[1,2]$ |

${x}_{low}\phantom{\rule{0.0pt}{0ex}}=\phantom{\rule{0.0pt}{0ex}}[0,2\pi ]$, ${x}_{high}\phantom{\rule{0.0pt}{0ex}}=\phantom{\rule{0.0pt}{0ex}}[{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}0.5\pi ,{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}2\pi ]$ | |||

High frequency | $a=[0.1s,0.1+\frac{s}{2f}]$, $f=[1.5,7]$ | ||

${x}_{low}\phantom{\rule{0.0pt}{0ex}}=\phantom{\rule{0.0pt}{0ex}}[0,2\pi ]$, ${x}_{high}\phantom{\rule{0.0pt}{0ex}}=\phantom{\rule{0.0pt}{0ex}}[{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}0.5\pi ,{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}2\pi ]$ | |||

Curved Waves | Quadratic | $f\left(x\right)=q\xb7{x}^{2}+$ | $q=[-10,15]$, $a=[0.05s,0.15s]$, $f=[0.5,2]$ |

$a\xb7sin(2\xb7\pi \xb7f\xb7x)$ | ${x}_{low}=[-3,3]$, ${x}_{high}=[{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}2,{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}6]$ | ||

Cubic | $f\left(x\right)=c\xb7{x}^{3}+q\xb7{x}^{2}+$ | $c=[-15,10]$, $q=[-15,30]$ | |

$a\xb7sin(2\xb7\pi \xb7f\xb7x)$ | $a=[0.1s,0.3s]$, $f=[0.5,2]$ | ||

${x}_{low}=[-3,3]$, ${x}_{high}=[{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}2,{x}_{low}\phantom{\rule{0.0pt}{0ex}}+\phantom{\rule{0.0pt}{0ex}}6]$ | |||

Bell curves | Gaussian | $f\left(x\right)=a\phantom{\rule{3.33333pt}{0ex}}\xb7{e}^{-\frac{{x}^{2}}{2{\sigma}^{2}}}$ | $a=[0.2s,s]$, $\sigma =[0.2,3]$ |

${x}_{low}=[-5,1]$, ${x}_{high}=[1,-5]$ | |||

G. inverted | $a=[-0.2s,-s]$, $\sigma =[0.2,3]$ | ||

${x}_{low}=[-5,1]$, ${x}_{high}=[1,-5]$ | |||

Polynomials | $f\left(x\right)=a\xb716\xb7$ | $a=[0.2s,s]$ | |

$({x}^{4}-2\xb7{x}^{3}+{x}^{2})$ | ${x}_{low}=0$, ${x}_{high}=1$ | ||

P. inverted | $a=[-0.2s,-s]$ | ||

${x}_{low}=0$, ${x}_{high}=1$ | |||

Bifurcations | Combination of all |

Inner Diameter in mm | Wall Thickness in mm |
---|---|

1 | 1 |

2 | 0.5 |

3 | 0.5 |

4 | 0.5 |

5 | 0.5 |

6 | 1 |

7 | 1 |

8 | 1 |

**Table 3.**Mean relative error of the discrete length measurement with no reconstruction compared to the continuous ground truth [25].

Function Type | Relative Error | Quantity in the Set |
---|---|---|

Straight lines | 2.3% | 184 |

Parabolas | 7.2% | 312 |

Polynomials | 6.3% | 168 |

Sinusoids | 6.8% | 132 |

Curved Waves | 7.6% | 102 |

Bell curves | 8.5% | 108 |

Bifurcations | 6.7% | 198 |

Mean of all | 6.3% |

**Table 4.**Mean relative error of the length measurement methods in % after centerline extraction by erosion compared to the continuous ground truth. Significance was proven for the reduction of the error compared to the discrete length for each function type ${}^{****}$ p < 0.0001.

Discrete μ ± σ | Bézier μ ± σ | Polynomial μ ± σ | |
---|---|---|---|

Straight lines | 5.2 ± 3.8 | 2.6 ± 1.8 ${}^{****}$ | 2.1 ± 1.6 ${}^{****}$ |

Parabolas | 7.5 ± 4.0 | 2.4 ± 1.5 ${}^{****}$ | 1.9 ± 1.3 ${}^{****}$ |

Polynomials | 6.5 ± 4.8 | 3.9 ± 6.2 ${}^{****}$ | 7.2 ± 7.5 ${}^{****}$ |

Sinusoids | 8.5 ± 5.8 | 4.2 ± 7.8 ${}^{****}$ | 6.6 ± 12.6 ${}^{****}$ |

Curved waves | 7.6 ± 2.6 | 2.5 ± 1.4 ${}^{****}$ | 2.6 ± 2.2 ${}^{****}$ |

Bell curves | 10.4 ± 3.2 | 3.1 ± 1.7 ${}^{****}$ | 1.9 ± 1.4 ${}^{****}$ |

Bifurcations | 6.1 ± 5.4 | 2.1 ± 5.3 ${}^{****}$ | 2.6 ± 5.1 ${}^{****}$ |

Mean of all | 7.0 ± 4.8 | 2.7 ± 4.5 ${}^{****}$ | 3.2 ± 5.8 ${}^{****}$ |

**Table 5.**Mean relative error of the length measurement methods in % after centerline extraction by Voronoi diagrams compared to the continuous ground truth. Significance was proven for the reduction of the error compared to the discrete length for each function type ${}^{****}$ p < 0.0001.

Discrete μ ± σ | Bézier μ ± σ | Polynomial μ ± σ | |
---|---|---|---|

Straight lines | 5.1 ± 5.8 | 3.3 ± 5.8 ${}^{****}$ | 3.2 ± 6.0 ${}^{****}$ |

Parabolas | 8.6 ± 7.0 | 4.2 ± 8.5 ${}^{****}$ | 4.5 ± 8.7 ${}^{****}$ |

Polynomials | 9.1 ± 8.5 | 7.7 ± 11.0 ${}^{****}$ | 11.2 ± 11.3 ${}^{****}$ |

Sinusoids | 8.6 ± 5.0 | 5.6 ± 7.8 ${}^{****}$ | 8.9 ± 12.8 ${}^{****}$ |

Curved waves | 8.3 ± 4.5 | 4.5 ± 7.0 ${}^{****}$ | 5.7 ± 7.5 ${}^{****}$ |

Bell curves | 8.2 ± 4.2 | 3.4 ± 5.1 ${}^{****}$ | 3.4 ± 5.6 ${}^{****}$ |

Bifurcations | 7.9 ± 11.2 | 4.6 ± 11.6 ${}^{****}$ | 5.9 ± 11.2 ${}^{****}$ |

Mean of all | 7.9 ± 8.1 | 4.7 ± 9.2 ${}^{****}$ | 5.9 ± 10.0 ${}^{****}$ |

Erosion | Voronoi | |
---|---|---|

Bézier curve | 13.08 s | 21.09 s |

Polynomial approximation | 2.33 s | 10.34 s |

**Table 7.**Mean relative error of the length measurement methods in % after the centerline extraction by both methods and the spatial interpolation compared to the ground truth for the experimental data set (mean also over all angles). Significance was proven for the reduction of the error compared to the discrete length for each function type ${}^{****}$ p < 0.0001. Angle resolved relative errors are given in Table 8 and Table 9.

Discrete μ ± σ | Bézier μ ± σ | Polynomial μ ± σ | |
---|---|---|---|

Erosion | 4.7 ± 3.0 | 1.9 ± 1.3 ${}^{****}$ | 1.6 ± 1.6 ${}^{****}$ |

Voronoi | 5.0 ± 3.0 | 2.0 ± 1.3 ${}^{****}$ | 1.5 ± 1.5 ${}^{****}$ |

**Table 8.**Mean relative error of the length measurement methods in % after the centerline extraction by erosion and the spatial interpolation compared to the ground truth for the experimental data set. Significance was proven for the reduction of the error compared to the discrete length as indicated ${}^{*}$ p < 0.05, ${}^{***}$p < 0.001, ${}^{****}$ p < 0.0001.

Discrete | Bézier | Polynomial | |
---|---|---|---|

Angle | μ ± σ | μ ± σ | μ ± σ |

0° | 2.4 ± 1.3 | 1.8 ±1.5 ${}^{*}$ | 1.5 ± 1.5 ${}^{***}$ |

15° | 7.0 ± 2.4 | 2.3 ± 1.0 ${}^{****}$ | 1.7 ± 1.6 ${}^{****}$ |

30° | 6.6 ± 2.3 | 1.9 ± 1.3 ${}^{****}$ | 1.6 ± 1.6 ${}^{****}$ |

45° | 2.2 ± 1.7 | 1.6 ± 1.6 ${}^{*}$ | 1.6 ± 1.7 ${}^{*}$ |

60° | 7.0 ± 2.3 | 1.8 ± 1.3 ${}^{****}$ | 1.4 ± 1.6 ${}^{****}$ |

75° | 5.8 ± 2.6 | 2.1 ± 1.2 ${}^{****}$ | 1.9 ± 1.7 ${}^{****}$ |

90° | 1.9 ± 1.3 | 1.7 ± 1.6 | 1.6 ± 1.7 |

**Table 9.**Mean relative error of the length measurement methods in % after the centerline extraction by Voronoi diagrams and the spatial interpolation compared to the ground truth for the experimental data set. Significance was proven for the reduction of the error compared to the discrete length as indicated ${}^{*}$p < 0.05, ${}^{**}$p < 0.01, ${}^{***}$p < 0.001, ${}^{****}$ p < 0.0001.

Discrete | Bézier | Polynomial | |
---|---|---|---|

Angle | μ ± σ | μ ± σ | μ ± σ |

0° | 2.4 ± 1.6 | 1.8 ±1.3 | 1.5 ± 1.4 ${}^{*}$ |

15° | 7.4 ± 2.3 | 2.6 ± 1.0 ${}^{****}$ | 1.6 ± 1.2 ${}^{****}$ |

30° | 6.8 ± 2.2 | 2.1 ± 1.3 ${}^{****}$ | 1.7 ± 1.6 ${}^{****}$ |

45° | 2.5 ± 1.5 | 1.6 ± 1.5 ${}^{**}$ | 1.5 ± 1.6 ${}^{***}$ |

60° | 7.1 ± 2.2 | 1.9 ± 1.1 ${}^{****}$ | 1.5 ± 1.5 ${}^{****}$ |

75° | 6.4 ± 2.2 | 2.3 ± 1.0 ${}^{****}$ | 1.5 ± 1.6 ${}^{****}$ |

90° | 2.3 ± 1.7 | 1.7 ± 1.5 | 1.5 ± 1.6 ${}^{**}$ |

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Naber, A.; Berwanger, D.; Nahm, W. Geodesic Length Measurement in Medical Images: Effect of the Discretization by the Camera Chip and Quantitative Assessment of Error Reduction Methods. *Photonics* **2020**, *7*, 70.
https://doi.org/10.3390/photonics7030070

**AMA Style**

Naber A, Berwanger D, Nahm W. Geodesic Length Measurement in Medical Images: Effect of the Discretization by the Camera Chip and Quantitative Assessment of Error Reduction Methods. *Photonics*. 2020; 7(3):70.
https://doi.org/10.3390/photonics7030070

**Chicago/Turabian Style**

Naber, Ady, Daniel Berwanger, and Werner Nahm. 2020. "Geodesic Length Measurement in Medical Images: Effect of the Discretization by the Camera Chip and Quantitative Assessment of Error Reduction Methods" *Photonics* 7, no. 3: 70.
https://doi.org/10.3390/photonics7030070