minimize Header

Prof. Dr.-Ing. Markus Mayer

Faculty of Computer Science

Professors

Professor

Study coordinator and advisor (from SS25 on) for the Bachelor study course “Artificial Intelligence” (in German)

K 006

0991/3615-8324

consulting time

In case my office door is open: Come in and ask if I currently have time! If I'm not present, please write me a mail to schedule an appointment.

publications

Sortierung:

Contribution

Markus Mayer
A. Borsdorf
H. Köstler
J. Hornegger
U. Rüde

Nonlinear Diffusion Noise Reduction in CT Using Correlation Analysis.

In: Proceedings of the 3rd Russian-Bavarian Conference on Biomedical Engineering.

(2007)

Contribution

Markus Mayer
A. Borsdorf
H. Köstler
J. Hornegger
U. Rüde

Nonlinear Diffusion vs. Wavelet Based Noise Reduction in CT Using Correlation Analysis.

In: Proceedings of Vision, Modeling, and Visualization (VMV) 2007. pg. 223-232

(2007)

Contribution

Markus Mayer
R.-P. Tornow
J. Hornegger
F. Kruse

Fuzzy C-means clustering for retinal layer segmentation on high resolution OCT images.

In: Proceedings of the 19th Biennial International EURASIP Conference (BIOSIGNAL).

(2008)

Journal article

Markus Mayer
R. Tornow
R. Bock
J. Hornegger
F. Kruse

Automatic nerve fiber layer segmentation and geometry correction on spectral domain OCT images using fuzzy C-means clustering.

In: Investigative Ophthalmology & Visual Science (IVOS) vol. 49 pg. 1880

(2008)

Contribution

Markus Mayer
R. Tornow
J. Bock
J. Hornegger
F. Kruse

Automatic nerve fiber layer segmentation and geometry correction on spectral domain OCT images using fuzzy C-means clustering.

In: Proceedings of Investigative Ophthalmology & Visual Science (ARVO Annual Meeting) 2008. pg. 1880

(2008)

Contribution

M. Wagner
A. Borsdorf
Markus Mayer
R. Tornow

Wavelet based approach to multiple-frame denoising of OCT images.

In: Proceedings of the 5th Russian-Bavarian Conference on Biomedical Engineering. pg. 67-69

Eds.:
H. Feußner

München

(2009)

Contribution

Markus Mayer
J. Hornegger
C. Mardin
F. Kruse
R. Tornow

Automated glaucoma classification using nerve fiber layer segmentations on circular spectral domain OCT B-scans.

In: Proceedings of Investigative Ophthalmology & Visual Science (ARVO Annual Meeting) 2009. (50) pg. 1101

(2009)

Contribution

M. Kraus
Markus Mayer
R. Bock
B. Potsaid
V. Manjunath
J. Duker
J. Hornegger
J. Fujimoto

Motion artifact correction in OCT volume scans using image registration.

In: Proceedings of Investigative Ophthalmology & Visual Science (ARVO Annual Meeting) 2010. (51) pg. 4405

(2010)

Contribution

Markus Mayer
M. Wagner
J. Hornegger
R. Tornow

Wavelet denoising of multiple-frame OCT data enhanced by a correlation analysis.

In: Proceedings of Investigative Ophthalmology & Visual Science (ARVO Annual Meeting) 2010. (51) pg. 1777

(2010)

Journal article

Markus Mayer
J. Hornegger
C. Mardin
R. Tornow

Retinal Nerve Fiber Layer Segmentation on FD-OCT Scans of Normal Subjects and Glaucoma Patients.

In: Biomedical Optics Express vol. 1 pg. 1358-1383

08.11.2010 (2010)

DOI: 10.1364/BOE.1.001358

show abstract

Automated measurements of the retinal nerve fiber layer thickness on circular OCT B-Scans provide physicians additional parameters for glaucoma diagnosis. We propose a novel retinal nerve fiber layer segmentation algorithm for frequency domain data that can be applied on scans from both normal healthy subjects, as well as glaucoma patients, using the same set of parameters. In addition, the algorithm remains almost unaffected by image quality. The main part of the segmentation process is based on the minimization of an energy function consisting of gradient and local smoothing terms. A quantitative evaluation comparing the automated segmentation results to manually corrected segmentations from three reviewers is performed. A total of 72 scans from glaucoma patients and 132 scans from normal subjects, all from different persons, composed the database for the evaluation of the segmentation algorithm. A mean absolute error per A-Scan of 2.9 µm was achieved on glaucomatous eyes, and 3.6 µm on healthy eyes. The mean absolute segmentation error over all A-Scans lies below 10 µm on 95.1% of the images. Thus our approach provides a reliable tool for extracting diagnostic relevant parameters from OCT B-Scans for glaucoma diagnosis.

Contribution

Markus Mayer
J. Hornegger
C. Mardin
R. Tornow

Retinal Layer Segmentation on OCT-Volume Scans of Normal and Glaucomatous Eyes.

In: Proceedings of Investigative Ophthalmology & Visual Science (ARVO Annual Meeting) 2011. (52) pg. 3669

(2011)

Contribution

R. Lämmer
R.-P. Tornow
Markus Mayer
F. Horn
F. Kruse
C. Mardin

Enhanced Depth Imaging Optical Coherence Tomography of the Choroid-Influence of Age and Glaucomatous Damage.

In: Proceedings of Investigative Ophthalmology & Visual Science (ARVO Annual Meeting) 2011. (52) pg. 3500

(2011)

Journal article

R. Tornow
W. Schrems
D. Bendschneider
F. Horn
Markus Mayer
C. Mardin
R. Lämmer

Atypical retardation patterns in scanning laser polarimetry are associated with low peripapillary choroidal thickness.

In: Investigative Ophthalmology & Visual Science (IVOS) vol. 52 pg. 7523-8

29.09.2011 (2011)

DOI: 10.1167/iovs.11-7557

show abstract

Purpose Scanning laser polarimetry (SLP) results can be affected by an atypical retardation pattern (ARP). One reason for an ARP is the birefringence of the sclera. The purpose of this study was to investigate the influence of the peripapillary choroidal thickness (pChTh) on the occurrence of ARP. Methods One hundred ten healthy subjects were investigated with SLP and spectral domain OCT. pChTh was measured in B-scan images at 768 positions using semiautomatic software. Values were averaged to 32 sectors and the total peripapillary mean. Subjects were divided into four groups according to the typical scan score (TSS) provided by the GDxVCC: group 1 TSS, 100; group 2 TSS, 90-99; group 3 TSS, 80-89; group 4 TSS, <80. Results Mean pChTh (± SD) in 110 healthy subjects was 141 μm (±49 μm). There was a significant correlation between pChTh and TSS (r = 0.608; P < 0.001). In TSS groups 1 to 4, mean pChTh was 168 μm (±38 μm), 148 μm (± 48 μm), 119 μm (±35 μm), and 92 (±42 μm). Mean pChTh of TSS groups 3 and 4 was significantly lower than that of TSS group 1 (P < 0.001). Conclusions Low values of TSS resulting from the appearance of ARP in SLP are associated with low peripapillary choroidal thickness. Reduced choroidal thickness may result in an increased amount of confounding light getting to the SLP light detectors.

Contribution

T. Köhler
J. Hornegger
Markus Mayer
G. Michelson

Quality-Guided Denoising for Low-Cost Fundus Imaging.

In: Bildverarbeitung für die Medizin 2012. (Informatik aktuell (INFORMAT)) pg. 292-297

Eds.:
T. Tolxdorff
H. Handels
H.-P. Meinzer
T. Deserno

Springer Berlin Heidelberg

(2012)

Contribution

S. Chitchian
Markus Mayer
A. Boretsky
F. van Kuijk
M. Motamedi

Complex Wavelet Denoising of Retinal OCT Imaging.

In: Proceedings of Investigative Ophthalmology & Visual Science (ARVO Annual Meeting) 2012. (53) pg. 3124

(2012)

Journal article

L. Balk
Markus Mayer
B. Uitdehaag
A. Petzold

Comparison of two retinal nerve fibre layer segmentation algorithms for the Heidelberg Spectralis.

In: Multiple Sclerosis Journal vol. 18 pg. 261-262

(2012)

Journal article

Markus Mayer
A. Borsdorf
M. Wagner
J. Hornegger
C. Mardin
R. Tornow

Wavelet denoising of multiframe optical coherence tomography data.

In: Biomedical Optics Express vol. 3 pg. 572-89

22.02.2012 (2012)

DOI: 10.1364/BOE.3.000572

show abstract

We introduce a novel speckle noise reduction algorithm for OCT images. Contrary to present approaches, the algorithm does not rely on simple averaging of multiple image frames or denoising on the final averaged image. Instead it uses wavelet decompositions of the single frames for a local noise and structure estimation. Based on this analysis, the wavelet detail coefficients are weighted, averaged and reconstructed. At a signal-to-noise gain at about 100% we observe only a minor sharpness decrease, as measured by a full-width-half-maximum reduction of 10.5%. While a similar signal-to-noise gain would require averaging of 29 frames, we achieve this result using only 8 frames as input to the algorithm. A possible application of the proposed algorithm is preprocessing in retinal structure segmentation algorithms, to allow a better differentiation between real tissue information and unwanted speckle noise.

Journal article

S. Chitchian
Markus Mayer
A. Boretsky
F. van Kuijk
M. Motamedi

Retinal optical coherence tomography image enhancement via shrinkage denoising using double-density dual-tree complex wavelet transform.

In: Journal of Biomedical Optics vol. 17 pg. 116009

(2012)

DOI: 10.1117/1.JBO.17.11.116009

show abstract

Image enhancement of retinal structures, in optical coherence tomography (OCT) scans through denoising, has the potential to aid in the diagnosis of several eye diseases. In this paper, a locally adaptive denoising algorithm using double-density dual-tree complex wavelet transform, a combination of the double-density wavelet transform and the dual-tree complex wavelet transform, is applied to reduce speckle noise in OCT images of the retina. The algorithm overcomes the limitations of commonly used multiple frame averaging technique, namely the limited number of frames that can be recorded due to eye movements, by providing a comparable image quality in significantly less acquisition time equal to an order of magnitude less time compared to the averaging method. In addition, improvements of image quality metrics and 5 dB increase in the signal-to-noise ratio are attained.

Contribution

I. Moupagiatzis
Markus Mayer
J. Hornegger
R. Tornow
C. Mardin

Application of morphological operators for optic nerve head segmentation in optical coherence tomography images.

In: Proceedings of the 2012 19th International Conference on Systems, Signals and Image Processing (IWSSIP).

(2012)

Journal article

M. Kraus
B. Potsaid
Markus Mayer
R. Bock
B. Baumann
J. Liu
J. Hornegger
J. Fujimoto

Motion correction in optical coherence tomography volumes on a per A-scan basis using orthogonal scan patterns.

In: Biomedical Optics Express vol. 3 pg. 1182-99

03.05.2012 (2012)

DOI: 10.1364/BOE.3.001182

show abstract

High speed Optical Coherence Tomography (OCT) has made it possible to rapidly capture densely sampled 3D volume data. One key application is the acquisition of high quality in vivo volumetric data sets of the human retina. Since the volume is acquired in a few seconds, eye movement during the scan process leads to distortion, which limits the accuracy of quantitative measurements using 3D OCT data. In this paper, we present a novel software based method to correct motion artifacts in OCT raster scans. Motion compensation is performed retrospectively using image registration algorithms on the OCT data sets themselves. Multiple, successively acquired volume scans with orthogonal fast scan directions are registered retrospectively in order to estimate and correct eye motion. Registration is performed by optimizing a large scale numerical problem as given by a global objective function using one dense displacement field for each input volume and special regularization based on the time structure of the acquisition process. After optimization, each volume is undistorted and a single merged volume is constructed that has superior signal quality compared to the input volumes. Experiments were performed using 3D OCT data from the macula and optic nerve head acquired with a high-speed ultra-high resolution 850 nm spectral OCT as well as wide field data acquired with a 1050 nm swept source OCT instrument. Evaluation of registration performance and result stability as well as visual inspection shows that the algorithm can correct for motion in all three dimensions and on a per A-scan basis. Corrected volumes do not show visible motion artifacts. In addition, merging multiple motion corrected and registered volumes leads to improved signal quality. These results demonstrate that motion correction and merging improves image quality and should also improve morphometric measurement accuracy from volumetric OCT data.

Contribution

S. Chitchian
Markus Mayer
A. Boretsky
F. van Kuijk
M. Motamedi

Retinal OCT Image Enhancement via Wavelet Denoising.

In: Biomedical Optics and 3-D Imaging. pg. BTu3A.73

OSA Washington, D.C.

(2012)

DOI: 10.1364/BIOMED.2012.BTu3A.73

Contribution

J. Fenolland
C. Boucher
Markus Mayer
W. Rostene
C. Baudouin
A. Denoyer

Developments in Optical Coherence Tomography Imaging in a Rat Model of Glaucoma.

In: Proceedings of Investigative Ophthalmology & Visual Science (ARVO Annual Meeting) 2013. (54) pg. 4833

(2013)

Journal article

L. Balk
Markus Mayer
B. Uitdehaag
A. Petzold

Physiological variation of segmented OCT retinal layer thicknesses is short-lasting.

In: Journal of Neurology vol. 260 pg. 3109-14

08.10.2013 (2013)

DOI: 10.1007/s00415-013-7097-6

show abstract

The application of spectral domain optical coherence tomography as a surrogate for neurodegeneration in a range of neurological disorders demands better understanding of the physiological variation of retinal layer thicknesses, which may mask any value of this emerging outcome measure. A prospective study compared retinal layer thicknesses between control subjects (n = 15) and runners (n = 27) participating in a 10-km charity run. Three scans were performed using an eye-tracking function (EBF) and automated scan registration for optimal precision at (1) baseline, (2) directly after the run, and (3) following a rehydration period. Retinal layer segmentation was performed with suppression of axial retinal vessel signal artifacts. Following the run, there was an increase in the relative retinal nerve fibre layer (p = 0.018), the combined inner plexiform/ganglion cell layer (p = 0.038), and the outer nuclear layer (p = 0.018) in runners compared to controls. The initial increase of thickness in the outer nuclear layer of runners (p < 0.0001) was likely related to (noncompliant) rehydration during exercise. Following a period of rest and rehydration, the difference in thickness change for all retinal layers, except the retinal nerve fibre layer (RNFL) (p < 0.05), disappeared between the two groups. There is a quantifiable change in the axial thickness of retinal layersthat which can be explained by an increase in the cellular volume. This effect may potentially be caused by H2O volume shifts.

Journal article

R. Kolar
R. Tornow
R. Laemmer
J. Odstrcilik
Markus Mayer
J. Gazarek
J. Jan
T. Kubena
P. Cernosek

Analysis of visual appearance of retinal nerve fibers in high resolution fundus images: a study on normal subjects.

In: Computational and Mathematical Methods in Medicine pg. 1-10

29.12.2013 (2013)

DOI: 10.1155/2013/134543

show abstract

The retinal ganglion axons are an important part of the visual system, which can be directly observed by fundus camera. The layer they form together inside the retina is the retinal nerve fiber layer (RNFL). This paper describes results of a texture RNFL analysis in color fundus photographs and compares these results with quantitative measurement of RNFL thickness obtained from optical coherence tomography on normal subjects. It is shown that local mean value, standard deviation, and Shannon entropy extracted from the green and blue channel of fundus images are correlated with corresponding RNFL thickness. The linear correlation coefficients achieved values 0.694, 0.547, and 0.512 for respective features measured on 439 retinal positions in the peripapillary area from 23 eyes of 15 different normal subjects.

Journal article

L. Balk
Markus Mayer
B. Uitdehaag
A. Petzold

Retinal hyperaemia-related blood vessel artifacts are relevant to automated OCT layer segmentation.

In: Journal of Neurology vol. 261 pg. 511-7

05.01.2014 (2014)

DOI: 10.1007/s00415-013-7226-2

show abstract

A frequently observed local measurement artifact with spectral domain OCT is caused by the void signal of the retinal vasculature. This study investigated the effect of suppression of blood vessel artifacts with and without retinal hyperaemia. Spectral domain OCT scans, centred on the optic nerve head, were performed in 46 healthy subjects (92 eyes). Baseline scans were made during rest, while for the follow-up scan, 23 subjects (50 %) performed strenuous physical exercise. Systemic and retinal hyperaemia were quantified. Quantification of retinal nerve fibre layer (RNFL) thickness was performed with and without suppression of retinal blood vessel artifacts. The potential systematic effect on RNFL thickness measurements was analysed using Bland-Altman plots. At baseline (no retinal hyperaemia), there was a systematic difference in RNFL thickness (3.4 μm, limits of agreement -0.9 to 7.7) with higher values if blood vessel artifacts were not suppressed. There was significant retinal hyperaemia in the exercise group (p < 0.0001). Baseline thickness increased from 93.18 to 93.83 μm (p < 0.05) in the exercise group using the algorithm with blood vessel artifact suppression, but no significant changes were observed using the algorithm without blood vessel artifact suppression. Retinal hyperaemia leads to blood vessel artifacts which are relevant to the precision of OCT layer segmentation algorithms. The two algorithms investigated in this study can not be used interchangeably. The algorithm with blood vessel artifact suppression was more sensitive in detecting small changes in RNFL thickness. This may be relevant for the use of OCT in a range of neurodegenerative diseases were only a small degree of retinal layer atrophy have been found so far.

Journal article

J. Odstrcilik
R. Kolar
R.-P. Tornow
J. Jan
A. Budai
Markus Mayer
M. Vodakova
R. Laemmer
M. Lamos
Z. Kuna
J. Gazarek
T. Kubena
P. Cernosek
M. Ronzhina

Thickness related textural properties of retinal nerve fiber layer in color fundus images.

In: Computerized Medical Imaging and Graphics: The Official Journal of the Computerized Medical Imaging Society vol. 38 pg. 508-16

21.05.2014 (2014)

DOI: 10.1016/j.compmedimag.2014.05.005

show abstract

Images of ocular fundus are routinely utilized in ophthalmology. Since an examination using fundus camera is relatively fast and cheap procedure, it can be used as a proper diagnostic tool for screening of retinal diseases such as the glaucoma. One of the glaucoma symptoms is progressive atrophy of the retinal nerve fiber layer (RNFL) resulting in variations of the RNFL thickness. Here, we introduce a novel approach to capture these variations using computer-aided analysis of the RNFL textural appearance in standard and easily available color fundus images. The proposed method uses the features based on Gaussian Markov random fields and local binary patterns, together with various regression models for prediction of the RNFL thickness. The approach allows description of the changes in RNFL texture, directly reflecting variations in the RNFL thickness. Evaluation of the method is carried out on 16 normal ("healthy") and 8 glaucomatous eyes. We achieved significant correlation (normals: ρ=0.72±0.14; p≪0.05, glaucomatous: ρ=0.58±0.10; p≪0.05) between values of the model predicted output and the RNFL thickness measured by optical coherence tomography, which is currently regarded as a standard glaucoma assessment device. The evaluation thus revealed good applicability of the proposed approach to measure possible RNFL thinning.

Patent

M. Kraus
B. Potsaid
J. Fujimoto
Markus Mayer
R. Bock
J. Hornegger

Method and apparatus for motion correction and image enhancement for optical coherence tomography.

Eds.:
Friedrich Alexander Universitaet Erlangen Nuernberg FAU
Massachusetts Institute of Technology

USA

23.02.2016 (2016)

Thesis

Markus Mayer

Automated Glaucoma Detection with Optical Coherence Tomography. Dissertationsschrift.

Friedrich-Alexander-Universität Erlangen-Nürnberg Erlangen-Nürnberg

25.07.2018 (2018)

core competencies

Medical image processing
Signal processing (audio & image processing and enhancement)
Pattern recognition (registration, segmentation)
Maschine learning (data mining, classification and regression)

Forschungs- und Lehrgebiete

Lectures:

Programming I & II
Statistics
Machine Learning
Project management
Industrial AI applications
AI
Theoretical Fundations of AI
Computer Sound

Vita

Resume:

2007: Completed computer science studies with side course musicology at the FAU Erlangen
2007-2012: PhD student at the Pattern Recognition Lab at the FAU Erlangen.
2012-2022: Researcher and developer at Arnold & Richter Cine Technik (Arri) in Munich. Last position: Senior image science engineer.
2022 (ongoing): Professor at the Deggendorf Institute of Technology

Accomplishments:

2008: International travel grant of the “Association for Research in Vision and Ophthalmology (ARVO) Annual Meeting 2008”
2011: Student award of the “Erlangen Graduate School of Advanced Optical Technologies (SAOT)“
2016: Patent granted (together with collegues from the LME Erlangen and the MIT Boston): “Method and apparatus for motion correction and image enhancement for optical coherence tomography”
2018: Finalized PhD: “Automated Glaucoma Detection with Optical Coherence Tomography”
Reviewer for multiple widely recognized magazines and organisations (a.o. IEEE Transaction on Medical Imaging, SPIE Journal of Biomedical Optics…)

Other

Open Bachelor- & Masterthesis topics:

AFFAIR - A Framework for arbitrary image registration
Neural Network Sound Synthesis (Follow up)
The “Mondrian” image compression (Follow up)
PROMUS - A Probabilistic Music Sequencer
LIREMA - Live Recorded Music Annotation
Audio Cycle Detection
Creating virtual instruments with Autoencoders
Room impule response simulation with neural networks
Replacing finite element simulations with neural networks

Currently running Bachelor- & Masterarbeiten (without company cooperation):

Optimization of noise for neural network confusion
Neural Network Sound Synthesis
Mondrian Image Compression
Generative Adversarial Networks for the Synthesis of Realistic Images from Sketches

These topics may not reflect the current state of what is available or in work. Contact me directly via mail if you want to know more. In general my topics are from the fields of:

Medical image processing
Machine learning and statistics in music
Interesting and creative challenges in signal- and image processing and machine learning