TY - JOUR
A1 - Schuldt, Thilo
A1 - Schubert, Christian
A1 - Krutzik, Markus
A1 - Bote, Lluis Gesa
A1 - Gaaloul, Naceur
A1 - Hartwig, Jonas
A1 - Ahlers, Holger
A1 - Herr, Waldemar
A1 - Posso-Trujillo, Katerine
A1 - Rudolph, Jan
A1 - Seidel, Stephan
A1 - Wendrich, Thijs
A1 - Ertmer, Wolfgang
A1 - Herrmann, Sven
A1 - Kubelka-Lange, André
A1 - Milke, Alexander
A1 - Rievers, Benny
A1 - Rocco, Emanuele
A1 - Hinton, Andrew
A1 - Bongs, Kai
A1 - Oswald, Markus
A1 - Franz, Matthias O.
A1 - Hauth, Matthias
A1 - Peters, Achim
A1 - Bawamia, Ahmad
A1 - Wicht, Andreas
A1 - Battelier, Baptiste
A1 - Bertoldi, Andrea
A1 - Bouyer, Philippe
A1 - Landragin, Arnaud
A1 - Massonnet, Didier
A1 - Léveque, Thomas
A1 - Wenzlawski, Andre
A1 - Hellmig, Ortwin
A1 - Windpassinger, Patrick
A1 - Sengstock, Klaus
A1 - von Klitzing, Wolf
A1 - Chaloner, Chris
A1 - Summers, David
A1 - Ireland, Philip
A1 - Mateos, Ignacio
A1 - Sopuerta, Carlos F.
A1 - Sorrentino, Fiodor
A1 - Tino, Guglielmo M.
A1 - Williams, Michael
A1 - Trenkel, Christian
A1 - Gerardi, Domenico
A1 - Chwalla, Michael
A1 - Burkhardt, Johannes
A1 - Johann, Ulrich
A1 - Heske, Astrid
A1 - Wille, Eric
A1 - Gehler, Martin
A1 - Cacciapuoti, Luigi
A1 - Gürlebeck, Norman
A1 - Braxmaier, Claus
A1 - Rasel, Ernst
T1 - Design of a dual species atom interferometer for space
JF - Experimental Astronomy
N2 - Atom interferometers have a multitude of proposed applications in space including precise measurements of the Earth's gravitational field, in navigation & ranging, and in fundamental physics such as tests of the weak equivalence principle (WEP) and gravitational wave detection. While atom interferometers are realized routinely in ground-based laboratories, current efforts aim at the development of a space compatible design optimized with respect to dimensions, weight, power consumption, mechanical robustness and radiation hardness. In this paper, we present a design of a high-sensitivity differential dual species 85Rb/87Rb atom interferometer for space, including physics package, laser system, electronics and software. The physics package comprises the atom source consisting of dispensers and a 2D magneto-optical trap (MOT), the science chamber with a 3D-MOT, a magnetic trap based on an atom chip and an optical dipole trap (ODT) used for Bose-Einstein condensate (BEC) creation and interferometry, the detection unit, the vacuum system for 10-11 mbar ultra-high vacuum generation, and the high-suppression factor magnetic shielding as well as the thermal control system.
The laser system is based on a hybrid approach using fiber-based telecom components and high-power laser diode technology and includes all laser sources for 2D-MOT, 3D-MOT, ODT, interferometry and detection. Manipulation and switching of the laser beams is carried out on an optical bench using Zerodur bonding technology. The instrument consists of 9 units with an overall mass of 221 kg, an average power consumption of 608 W (819 W peak), and a volume of 470 liters which would well fit on a satellite to be launched with a Soyuz rocket, as system studies have shown.
KW - Atom interferometer
KW - Space technology
KW - Equivalence principle test
KW - Bose-Einstein condensate
Y1 - 2015
U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bsz:kon4-opus4-4116
SN - 1572-9508
VL - 39
IS - 2
SP - 167
EP - 206
ER -
TY - GEN
A1 - Franz, Matthias O.
T1 - Volterra and Wiener series
T2 - Scholarpedia
N2 - Volterra and Wiener series are two classes of polynomial representations of nonlinear systems. They are perhaps the best understood and most widely used nonlinear system representations in signal processing and system identification. A Volterra or Wiener representation can be thought of as a natural extension of the classical linear system representation. In addition to the convolution of the input signal with the system's impulse response, the system representation includes a series of nonlinear terms that contain products of increasing order of the input signal with itself. It can be shown that these polynomial extension terms allow for representing a large class of nonlinear systems which basically encompasses all systems with scalar outputs that are time-invariant and have noninfinite memory.
KW - System identification
KW - Kernel based learning
Y1 - 2011
UR - http://www.scholarpedia.org/article/Volterra_and_Wiener_series
U6 - http://dx.doi.org/doi:10.4249/scholarpedia.11307
IS - 6(10):11307
ER -
TY - CHAP
A1 - Grunwald, Michael
A1 - Franz, Matthias O.
T1 - Wahrnehmungsorientierte optische Inspektion von texturierten Oberflächen
T2 - Informatik 2016 - Tagung vom 26.-30. September 2016, Klagenfurt
Y1 - 2016
UR - http://subs.emis.de/LNI/Proceedings/Proceedings259/1963.pdf
SN - 978-3-88579-653-4
SP - 1963
EP - 1968
ER -
TY - JOUR
A1 - Grunwald, Michael
A1 - Siebeck, Ulrike
A1 - Franz, Matthias O.
T1 - FishNet
BT - Automatisierte Erfassung von Fischbeständen für die Klimaforschung
JF - horizonte
Y1 - 2016
UR - https://www.koord.hs-mannheim.de/fileadmin/user_upload/projekte/koord/horizonte/h47_komplett.pdf
SN - 1432-9174
IS - 47
SP - 12
EP - 14
ER -
TY - CHAP
A1 - Grunwald, Michael
A1 - Gansloser, Jens
A1 - Franz, Matthias O.
T1 - Radiometric calibration of digital cameras using sparse Gaussian processes
T2 - 22. Workshop Farbbildverarbeitung : 29.-30. September 2016, Ilmenau
N2 - Digital cameras are used in a large variety of scientific and industrial applications. For most applications the acquired data should represent the real light intensity per pixel as accurately as possible. However, digital cameras are subject to different sources of noise which distort the resulting image. Noise includes photon noise, fixed pattern noise and read noise. The aim of the radiometric calibration is to improve the quality of the resulting images by reducing the influence of the different types of noise on the measured data. In this paper, a new approach for the radiometric calibration of digital cameras using sparse Gaussian process regression is presented. Gaussian process regression is a kernel based supervised machine learning technique. It is used to learn the response of a camera system from a set of training images to allow for the calibration of new images. Compared to the standard Gaussian process method or flat field correction our sparse approach allows for faster calibration and higher reconstruction quality.
Y1 - 2016
UR - https://www.researchgate.net/publication/310313872_Radiometric_calibration_of_digital_cameras_using_sparse_Gaussian_processes
SN - 978-3-00-053918-3
SP - 23
EP - 35
ER -
TY - CHAP
A1 - Grunwald, Michael
A1 - Müller, Jens
A1 - Schall, Martin
A1 - Franz, Matthias O.
T1 - Pixel-wise Hybrid Image Registration on Wood Decors
T2 - BW-CAR Symposium on Information and Communication Systems, SInCom 2015, 13. November 2015, Konstanz
N2 - The detection of differences between images of a printed reference and a reprinted wood decor often requires an initial image registration step. Depending on the digitalization method, the reprint will be displaced and rotated with respect to the reference. The aim of registration is to match the images as precisely as possible. In our approach, images are first matched globally by extracting feature points from both images and finding corresponding point pairs using the RANSAC algorithm. From these correspondences, we compute a global projective transformation between both images. In order to get a pixel-wise registration, we train a learning machine on the point correspondences found by RANSAC. The learning algorithm (in our case Gaussian process regression) is used to nonlinearly interpolate between the feature points which results in a high precision image registration method on wood decors.
Y1 - 2015
UR - https://opus.htwg-konstanz.de/frontdoor/index/index/docId/444
SN - 978-3-00-051859-1
SP - 24
EP - 29
ER -
TY - CHAP
A1 - Schall, Martin
A1 - Schambach, Marc-Peter
A1 - Franz, Matthias O.
T1 - Improving gradient-based LSTM training for offline handwriting recognition by careful selection of the optimization method
T2 - 3rd Baden-Württemberg Center of Applied Research Symposium on Information and Communication Systems - SInCom 2016 - Karlsruhe, December 2nd, 2016
N2 - Recent years have seen the proposal of several different gradient-based optimization methods for training artificial neural networks. Traditional methods include steepest descent with momentum, newer methods are based on per-parameter learning rates and some approximate Newton-step updates. This work contains the result of several experiments comparing different optimization methods. The experiments were targeted at offline handwriting recognition using hierarchical subsampling networks with recurrent LSTM layers. We present an overview of the used optimization methods, the results that were achieved and a discussion of why the methods lead to different results.
Y1 - 2016
U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bsz:ofb1-opus4-17866
UR - https://www.researchgate.net/publication/311439914_Improving_gradient-based_LSTM_training_for_offline_handwriting_recognition_by_careful_selection_of_the_optimization_method
SN - 978-3-943301-21-2
SP - 11
EP - 16
ER -
TY - GEN
A1 - Umlauf, Georg
A1 - Laube, Pascal
A1 - Franz, Matthias O.
A1 - Blume, Merlin
A1 - Caputo, Manuel
T1 - Using machine learning methods in geometric modeling
T2 - Presentation at Geometric Modelling, Interoperability and New Challenges, Schloss Dagstuhl - Leibniz-Zentrum für Informatik, 1 June 2017, Wadern, D
Y1 - 2017
UR - http://materials.dagstuhl.de/files/17/17221/17221.GeorgUmlauf2.Slides.pdf
ER -
TY - GEN
A1 - Scholten, Anja
A1 - Franz, Matthias O.
T1 - The impact of low water periods on inland navigation and mass cargo affine companies under climate change conditions
T2 - Zentralkommission für die Rheinschifffahrt (ZKR), Straßburg, 21.03.2017
N2 - Vortrag
Y1 - 2017
ER -
TY - GEN
A1 - Schall, Martin
A1 - Franz, Matthias O.
T1 - Segmentation-free multi-line text recognition using LSTM networks
T2 - Summerschool der Universität Konstanz, Gaschurn, Österreich, 2017
N2 - Vortrag
Y1 - 2017
ER -
TY - CHAP
A1 - Laube, Pascal
A1 - Franz, Matthias O.
A1 - Umlauf, Georg
T1 - Evaluation of features for SVM-based classification of geometric primitives in point clouds
T2 - Proceedings of the Fifteenth IAPR International Conference on Machine Vision Applications (MVA), 8-12 May 2017, Nagoya, Japan
N2 - In the reverse engineering process one has to classify parts of point clouds with the correct type of geometric primitive. Features based on different geometric properties like point relations, normals, and curvature information can be used, to train classifiers like Support Vector Machines (SVM). These geometric features are estimated in the local neighborhood of a point of the point cloud. The multitude of different features makes an in-depth comparison necessary. In this work we evaluate 23 features for the classification of geometric primitives in point clouds. Their performance is evaluated on SVMs when used to classify geometric primitives in simulated and real laser scanned point clouds. We also introduce a normalization of point cloud density to improve classification generalization.
Y1 - 2017
SN - 978-4-9011-2216-0
U6 - http://dx.doi.org/10.23919/MVA.2017.7986776
N1 - Volltextzugriff für Angehörige der Hochschule Konstanz möglich
SP - 59
EP - 62
ER -
TY - CHAP
A1 - Laube, Pascal
A1 - Franz, Matthias O.
A1 - Umlauf, Georg
T1 - Deep 3D
BT - machine learning for reconstruction and repair of 3D surfaces
T2 - GTC Europe 2017, NVIDIA GPU Technology Conference, 10.-12. Oktober 2017, ICM München
Y1 - 2017
UR - https://on-demand-gtc.gputechconf.com/gtcnew/sessionview.php?sessionName=23152-deep+3d+-+machine+learning+for+reconstruction+and+repair+of+3d+surfaces
ER -
TY - GEN
A1 - Laube, Pascal
A1 - Franz, Matthias O.
A1 - Umlauf, Georg
T1 - Deep learning for reconstruction of highly detailed surfaces
T2 - Summerschool der Universität Konstanz, Gaschurn, Österreich, 2017
N2 - Vortrag
Y1 - 2017
ER -
TY - CHAP
A1 - Schall, Martin
A1 - Sacha, Dominik
A1 - Stein, Manuel
A1 - Franz, Matthias O.
A1 - Keim, Daniel
T1 - Visualization-Assisted Development of Deep Learning Models in Offline Handwriting Recognition
T2 - Visualization in Data Science (VDS at IEEE VIS 2018), 22 October 2018, Berlin, Germany
N2 - Deep learning is a field of machine learning that has been the focus of active research and successful applications in recent years. Offline handwriting recognition is one of the research fields and applications were deep neural networks have shown high accuracy. Deep learning models and their training pipeline show a large amount of hyper-parameters in their data selection, transformation, network topology and training process that are sometimes interdependent. This increases the overall difficulty and time necessary for building and training a model for a specific data set and task at hand. This work proposes a novel visualization-assisted workflow that guides the model developer through the hyper-parameter search in order to identify relevant parameters and modify them in a meaningful way. This decreases the overall time necessary for building and training a model. The contributions of this work are a workflow for hyper-parameter search in offline handwriting recognition and a heat map based visualization technique for deep neural networks in multi-line offline handwriting recognition. This work applies to offline handwriting recognition, but the general workflow can possibly be adapted to other tasks as well.
Y1 - 2018
UR - https://www.researchgate.net/publication/328687407
ER -
TY - CHAP
A1 - Caputo, Manuel
A1 - Denker, Klaus
A1 - Franz, Matthias O.
A1 - Laube, Pascal
A1 - Umlauf, Georg
T1 - Support Vector Machines for Classification of Geometric Primitives in Point Clouds
T2 - Curves and Surfaces : 8th International Conference, Paris, France, June 12-18, 2014
N2 - Classification of point clouds by different types of geometric primitives is an essential part in the reconstruction process of CAD geometry. We use support vector machines (SVM) to label patches in point clouds with the class labels tori, ellipsoids, spheres, cones, cylinders or planes. For the classification features based on different geometric properties like point normals, angles, and principal curvatures are used. These geometric features are estimated in the local neighborhood of a point of the point cloud. Computing these geometric features for a random subset of the point cloud yields a feature distribution. Different features are combined for achieving best classification results. To minimize the time consuming training phase of SVMs, the geometric features are first evaluated using linear discriminant analysis (LDA).
LDA and SVM are machine learning approaches that require an initial training phase to allow for a subsequent automatic classification of a new data set. For the training phase point clouds are generated using a simulation of a laser scanning device. Additional noise based on an laser scanner error model is added to the point clouds. The resulting LDA and SVM classifiers are then used to classify geometric primitives in simulated and real laser scanned point clouds.
Compared to other approaches, where all known features are used for classification, we explicitly compare novel against known geometric features to prove their effectiveness.
Y1 - 2015
SN - 978-3-319-22804-4
U6 - http://dx.doi.org/10.1007/978-3-319-22804-4_7
N1 - Volltextzugriff für Hochschulangehörige möglich
SP - 80
EP - 95
PB - Springer
CY - Cham
ER -
TY - CHAP
A1 - Schall, Martin
A1 - Grunwald, Michael
A1 - Umlauf, Georg
A1 - Franz, Matthias O.
T1 - Radiometric calibration of digital cameras using Gaussian processes
T2 - Optical Sensors 2015, SPIE OPTICS + OPTOELECTRONICS
13-16 April 2015, Prague, Czech Republic ; Proceedings of SPIE
N2 - Digital cameras are subject to physical, electronic and optic effects that result in errors and noise in the image. These effects include for example a temperature dependent dark current, read noise, optical vignetting or different sensitivities of individual pixels. The task of a radiometric calibration is to reduce these errors in the image and thus improve the quality of the overall application. In this work we present an algorithm for radiometric calibration based on Gaussian processes. Gaussian processes are a regression method widely used in machine learning that is particularly useful in our context. Then Gaussian process regression is used to learn a temperature and exposure time dependent mapping from observed gray-scale values to true light intensities for each pixel. Regression models based on the characteristics of single pixels suffer from excessively high runtime and thus are unsuitable for many practical applications. In contrast, a single regression model for an entire image with high spatial resolution leads to a low quality radiometric calibration, which also limits its practical use. The proposed algorithm is predicated on a partitioning of the pixels such that each pixel partition can be represented by one single regression model without quality loss. Partitioning is done by extracting features from the characteristic of each pixel and using them for lexicographic sorting. Splitting the sorted data into partitions with equal size yields the final partitions, each of which is represented by the partition centers. An individual Gaussian process regression and model selection is done for each partition. Calibration is performed by interpolating the gray-scale value of each pixel with the regression model of the respective partition. The experimental comparison of the proposed approach to classical flat field calibration shows a consistently higher reconstruction quality for the same overall number of calibration frames.
Y1 - 2015
SN - 978-162841-627-5
U6 - http://dx.doi.org/10.1117/12.2178601
IS - Volume 9506
PB - SPIE
CY - Bellingham, Washington
ER -
TY - CHAP
A1 - Schall, Martin
A1 - Schambach, Marc-Peter
A1 - Franz, Matthias O.
T1 - Increasing robustness of handwriting recognition using character N-Gram decoding on large lexica
T2 - 12th IAPR Workshop on Document Analysis Systems (DAS),
11-14 April 2016
N2 - Offline handwriting recognition systems often include a decoding step, that is retrieving the most likely character sequence from the underlying machine learning algorithm. Decoding is sensitive to ranges of weakly predicted characters, caused e.g. by obstructions in the scanned document. We present a new algorithm for robust decoding of handwriting recognizer outputs using character n-grams. Multidimensional hierarchical subsampling artificial neural networks with Long-Short-Term-Memory cells have been successfully applied to offline handwriting recognition. Output activations from such networks, trained with Connectionist Temporal Classification, can be decoded with several different algorithms in order to retrieve the most likely literal string that it represents. We present a new algorithm for decoding the network output while restricting the possible strings to a large lexicon. The index used for this work is an n-gram index with tri-grams used for experimental comparisons. N-grams are extracted from the network output using a backtracking algorithm and each n-gram assigned a mean probability. The decoding result is obtained by intersecting the n-gram hit lists while calculating the total probability for each matched lexicon entry. We conclude with an experimental comparison of different decoding algorithms on a large lexicon.
KW - Probability
KW - Computational linguistics
KW - Document image processing
KW - Handwriting recognition
KW - Learning (artificial intelligence)
Y1 - 2016
U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bsz:kon4-opus4-9418
UR - https://ieeexplore.ieee.org/document/7490110
SN - 978-1-5090-1792-8
SP - 1
EP - 6
PB - IEEE
ER -
TY - JOUR
A1 - Aguilera, D.
A1 - Ahlers, Holger
A1 - Battelier, Baptiste
A1 - Bawamia, Ahmad
A1 - Bertoldi, Andrea
A1 - Bondarescu, R.
A1 - Bongs, Kai
A1 - Bouyer, Philippe
A1 - Braxmaier, Claus
A1 - Cacciapuoti, Luigi
A1 - Chaloner, Chris
A1 - Chwalla, Michael
A1 - Ertmer, Wolfgang
A1 - Franz, Matthias O.
A1 - Gaaloul, Naceur
A1 - Gehler, Martin
A1 - Gerardi, Domenico
A1 - Gesa, L.
A1 - Gürlebeck, Norman
A1 - Hartwig, Jonas
A1 - Hauth, Matthias
A1 - Hellmig, Ortwin
A1 - Herr, Waldemar
A1 - Herrmann, Sven
A1 - Heske, Astrid
A1 - Hinton, Andrew
A1 - Ireland, Philip
A1 - Jetzer, P.
A1 - Johann, Ulrich
A1 - Krutzik, Markus
A1 - Kubelka-Lange, André
A1 - Lämmerzahl, C.
A1 - Landragin, Arnaud
A1 - Lloro, I.
A1 - Massonnet, Didier
A1 - Mateos, Ignacio
A1 - Milke, Alexander
A1 - Nofrarias, M.
A1 - Oswald, Markus
A1 - Peters, Achim
A1 - Posso-Trujillo, Katerine
A1 - Rasel, Ernst
A1 - Rocco, Emanuele
A1 - Roura, A.
A1 - Rudolph, J.
A1 - Schleich, W.
A1 - Schubert, Christian
A1 - Schuldt, Thilo
A1 - Seidel, Stephan
A1 - Sengstock, Klaus
A1 - Sopuerta, Carlos F.
A1 - Sorrentino, Fiodor
A1 - Summers, David
A1 - Tino, Guglielmo M.
A1 - Trenkel, Christian
A1 - Uzunoglu, N.
A1 - von Klitzing, Wolf
A1 - Walser, R.
A1 - Wendrich, Thijs
A1 - Wenzlawski, Andre
A1 - Weßels, P.
A1 - Wicht, Andreas
A1 - Wille, Eric
A1 - Williams, Michael
A1 - Windpassinger, Patrick
A1 - Zahzam, N.
T1 - STE-QUEST - Test of the Universality of Free Fall Using Cold Atom Interferometry
JF - Classical and Quantum Gravity
N2 - In this paper, we report about the results of the phase A mission study of the atom
interferometer instrument covering the description of the main payload elements, the
atomic source concept, and the systematic error sources.
KW - Atom interferometry
KW - Equivalence principle
KW - Cold atoms
KW - Bose-Einstein condensate
KW - Microgravity
KW - Quantum gravity
KW - Space physics
Y1 - 2014
U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bsz:kon4-opus4-4157
SN - 0264-9381
VL - 31
IS - 11
ER -
TY - CHAP
A1 - Laube, Pascal
A1 - Franz, Matthias O.
A1 - Grunwald, Michael
A1 - Umlauf, Georg
T1 - Image Inpainting for High-Resolution Textures using CNN Texture Synthesis
BT - Short Paper
T2 - Computer Graphics & Visual Computing (CGVC) 2018,
13th - 14th September 2018, Swansea University, United Kingdom
N2 - Deep neural networks have been successfully applied to problems such as image segmentation, image super-resolution, coloration and image inpainting. In this work we propose the use of convolutional neural networks (CNN) for image inpainting of large regions in high-resolution textures. Due to limited computational resources processing high-resolution images with neural networks is still an open problem. Existing methods separate inpainting of global structure and the transfer of details, which leads to blurry results and loss of global coherence in the detail transfer step. Based on advances in texture synthesis using CNNs we propose patch-based image inpainting by a single network topology that is able to optimize for global as well as detail texture statistics. Our method is capable of filling large inpainting regions, oftentimes exceeding quality of comparable methods for images of high-resolution (2048x2048px). For reference patch look-up we propose to use the same summary statistics that are used in the inpainting process.
Y1 - 2018
UR - https://arxiv.org/abs/1712.03111v2
ET - Version 2
ER -
TY - JOUR
A1 - Laube, Pascal
A1 - Franz, Matthias O.
A1 - Umlauf, Georg
T1 - Learnt knot placement in B-spline curve approximation using support vector machines
JF - Computer Aided Geometric Design
N2 - Knot placement for curve approximation is a well known and yet open problem in geometric modeling. Selecting knot values that yield good approximations is a challenging task, based largely on heuristics and user experience. More advanced approaches range from parametric averaging to genetic algorithms.
In this paper, we propose to use Support Vector Machines (SVMs) to determine suitable knot vectors for B-spline curve approximation. The SVMs are trained to identify locations in a sequential point cloud where knot placement will improve the approximation error. After the training phase, the SVM can assign, to each point set location, a so-called score. This score is based on geometric and differential geometric features of points. It measures the quality of each location to be used as knots in the subsequent approximation. From these scores, the final knot vector can be constructed exploring the topography of the score-vector without the need for iteration or optimization in the approximation process. Knot vectors computed with our approach outperform state of the art methods and yield tighter approximations.
Y1 - 2018
U6 - http://dx.doi.org/10.1016/j.cagd.2018.03.019
SN - 0167-8396
VL - 62
SP - 104
EP - 116
ER -
TY - CHAP
A1 - Grunwald, Michael
A1 - Hermann, Matthias
A1 - Freiberg, Fabian
A1 - Laube, Pascal
A1 - Franz, Matthias O.
T1 - Optical surface inspection: A novelty detection approach based on CNN-encoded texture features
T2 - Applications of Digital Image Processing XLI,
19-23 August 2018, San Diego, California (Proceedings of SPIE Optical Engineering & Applications, Vol. 10752)
N2 - In inspection systems for textured surfaces, a reference texture is typically known before novel examples are inspected. Mostly, the reference is only available in a digital format. As a consequence, there is no dataset of defective examples available that could be used to train a classifier. We propose a texture model approach to novelty detection. The texture model uses features encoded by a convolutional neural network (CNN) trained on natural image data. The CNN activations represent the specific characteristics of the digital reference texture which are learned by a one-class classifier. We evaluate our novelty detector in a digital print inspection scenario. The inspection unit is based on a camera array and a flashing light illumination which allows for inline capturing of multichannel images at a high rate. In order to compare our results to manual inspection, we integrated our inspection unit into an industrial single-pass printing system.
Y1 - 2018
U6 - http://dx.doi.org/10.1117/12.2320657
ER -
TY - CHAP
A1 - Schall, Martin
A1 - Buehrig, Haiyan
A1 - Schambach, Marc-Peter
A1 - Franz, Matthias O.
T1 - LSTM Networks for Edit Distance Calculation with Exchangeable Dictionaries
T2 - 13th IAPR International Workshop on Document Analysis Systems, 24 - 27. April 2018, Vienna, Austria, (DAS 2018, Short Paper Booklet, Preprint)
Y1 - 2018
UR - https://das2018.cvl.tuwien.ac.at/media/filer_public/85/fd/85fd4698-040f-45f4-8fcc-56d66533b82d/das2018_short_papers.pdf
SP - 17
EP - 18
ER -
TY - CHAP
A1 - Schall, Martin
A1 - Schambach, Marc-Peter
A1 - Franz, Matthias O.
T1 - Multi-Dimensional Connectionist Classification
BT - Reading Text in One Step
T2 - 13th IAPR International Workshop on Document Analysis Systems, 24 - 27. April 2018, Vienna, Austria
Y1 - 2018
UR - https://ieeexplore.ieee.org/document/8395230
SN - 978-1-5386-3346-5
U6 - http://dx.doi.org/10.1109/DAS.2018.36
N1 - Volltextzugriff für Angehörige der Hochschule Konstanz möglich.
SP - 405
EP - 410
ER -
TY - CHAP
A1 - Laube, Pascal
A1 - Franz, Matthias O.
A1 - Umlauf, Georg
T1 - Deep Learning Parametrization for B-Spline Curve Approximation
T2 - International Conference on 3D Vision (3DV), 5-8 Sept. 2018, Verona, Italy
N2 - In this paper we present a method using deep learning to compute parametrizations for B-spline curve approximation. Existing methods consider the computation of parametric values and a knot vector as separate problems. We propose to train interdependent deep neural networks to predict parametric values and knots. We show that it is possible to include B-spline curve approximation directly into the neural network architecture. The resulting parametrizations yield tight approximations and are able to outperform state-of-the-art methods.
Y1 - 2018
U6 - http://dx.doi.org/10.1109/3DV.2018.00084
SN - 2475-7888
N1 - Volltextzugriff im Campusnetz der Hochschule Konstanz möglich (Datenbank IEEE Xplore)
SP - 691
EP - 699
PB - IEEE
ER -
TY - RPRT
A1 - Franz, Matthias O.
T1 - Fortbildungssemester in Rotorua, Neuseeland, Wintersemester 2018/19
BT - Erfahrungsbericht
Y1 - 2019
U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bsz:kon4-opus4-20852
ER -
TY - CHAP
A1 - Franz, Matthias O.
T1 - Machine learning as model of biological decision processes
T2 - University of Queensland, Brisbane, Australien, 2015
N2 - Vortrag
Y1 - 2015
ER -