8 Search Results for "Winter, Stefan"

Issue

DARTS, Volume 9, Issue 2

Special Issue of the 37th European Conference on Object-Oriented Programming (ECOOP 2023)

Editors: Hernán Ponce de León and Stefan Winter

Issue

DARTS, Volume 8, Issue 2

Special Issue of the 36th European Conference on Object-Oriented Programming (ECOOP 2022)

Editors: Alessandra Gorla and Stefan Winter

Document
DOI: 10.4230/LITES.8.1.1
Susceptibility to Image Resolution in Face Recognition and Training Strategies to Enhance Robustness

Authors: Martin Knoche, Stefan Hörmann, and Gerhard Rigoll

Published in: LITES, Volume 8, Issue 1 (2022): Special Issue on Embedded Systems for Computer Vision. Leibniz Transactions on Embedded Systems, Volume 8, Issue 1

Abstract
Many face recognition approaches expect the input images to have similar image resolution. However, in real-world applications, the image resolution varies due to different image capture mechanisms or sources, affecting the performance of face recognition systems. This work first analyzes the image resolution susceptibility of modern face recognition. Face verification on the very popular LFW dataset drops from 99.23% accuracy to almost 55% when image dimensions of both images are reduced to arguable very poor resolution. With cross-resolution image pairs (one HR and one LR image), face verification accuracy is even worse. This characteristic is investigated more in-depth by analyzing the feature distances utilized for face verification. To increase the robustness, we propose two training strategies applied to a state-of-the-art face recognition model: 1) Training with 50% low resolution images within each batch and 2) using the cosine distance loss between high and low resolution features in a siamese network structure. Both methods significantly boost face verification accuracy for matching training and testing image resolutions. Training a network with different resolutions simultaneously instead of adding only one specific low resolution showed improvements across all resolutions and made a single model applicable to unknown resolutions. However, models trained for one particular low resolution perform better when using the exact resolution for testing. We improve the face verification accuracy from 96.86% to 97.72% on the popular LFW database with uniformly distributed image dimensions between 112 × 112 px and 5 × 5 px. Our approaches improve face verification accuracy even more from 77.56% to 87.17% for distributions focusing on lower images resolutions. Lastly, we propose specific image dimension sets focusing on high, mid, and low resolution for five well-known datasets to benchmark face verification accuracy in cross-resolution scenarios.
Cite as

Martin Knoche, Stefan Hörmann, and Gerhard Rigoll. Susceptibility to Image Resolution in Face Recognition and Training Strategies to Enhance Robustness. In LITES, Volume 8, Issue 1 (2022): Special Issue on Embedded Systems for Computer Vision. Leibniz Transactions on Embedded Systems, Volume 8, Issue 1, pp. 01:1-01:20, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2022)

Copy BibTex To Clipboard

@Article{knoche_et_al:LITES.8.1.1,
  author =	{Knoche, Martin and H\"{o}rmann, Stefan and Rigoll, Gerhard},
  title =	{{Susceptibility to Image Resolution in Face Recognition and Training Strategies to Enhance Robustness}},
  journal =	{Leibniz Transactions on Embedded Systems},
  pages =	{01:1--01:20},
  ISSN =	{2199-2002},
  year =	{2022},
  volume =	{8},
  number =	{1},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LITES.8.1.1},
  doi =		{10.4230/LITES.8.1.1},
  annote =	{Keywords: recognition, resolution, cross, face, identification}
}
Document
DOI: 10.4230/LITES.8.1.3
HW-Flow: A Multi-Abstraction Level HW-CNN Codesign Pruning Methodology

Authors: Manoj-Rohit Vemparala, Nael Fasfous, Alexander Frickenstein, Emanuele Valpreda, Manfredi Camalleri, Qi Zhao, Christian Unger, Naveen-Shankar Nagaraja, Maurizio Martina, and Walter Stechele

Published in: LITES, Volume 8, Issue 1 (2022): Special Issue on Embedded Systems for Computer Vision. Leibniz Transactions on Embedded Systems, Volume 8, Issue 1

Abstract
Convolutional neural networks (CNNs) have produced unprecedented accuracy for many computer vision problems in the recent past. In power and compute-constrained embedded platforms, deploying modern CNNs can present many challenges. Most CNN architectures do not run in real-time due to the high number of computational operations involved during the inference phase. This emphasizes the role of CNN optimization techniques in early design space exploration. To estimate their efficacy in satisfying the target constraints, existing techniques are either hardware (HW) agnostic, pseudo-HW-aware by considering parameter and operation counts, or HW-aware through inflexible hardware-in-the-loop (HIL) setups. In this work, we introduce HW-Flow, a framework for optimizing and exploring CNN models based on three levels of hardware abstraction: Coarse, Mid and Fine. Through these levels, CNN design and optimization can be iteratively refined towards efficient execution on the target hardware platform. We present HW-Flow in the context of CNN pruning by augmenting a reinforcement learning agent with key metrics to understand the influence of its pruning actions on the inference hardware. With 2× reduction in energy and latency, we prune ResNet56, ResNet50, and DeepLabv3 with minimal accuracy degradation on the CIFAR-10, ImageNet, and CityScapes datasets, respectively.
Cite as

Manoj-Rohit Vemparala, Nael Fasfous, Alexander Frickenstein, Emanuele Valpreda, Manfredi Camalleri, Qi Zhao, Christian Unger, Naveen-Shankar Nagaraja, Maurizio Martina, and Walter Stechele. HW-Flow: A Multi-Abstraction Level HW-CNN Codesign Pruning Methodology. In LITES, Volume 8, Issue 1 (2022): Special Issue on Embedded Systems for Computer Vision. Leibniz Transactions on Embedded Systems, Volume 8, Issue 1, pp. 03:1-03:30, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2022)

Copy BibTex To Clipboard

@Article{vemparala_et_al:LITES.8.1.3,
  author =	{Vemparala, Manoj-Rohit and Fasfous, Nael and Frickenstein, Alexander and Valpreda, Emanuele and Camalleri, Manfredi and Zhao, Qi and Unger, Christian and Nagaraja, Naveen-Shankar and Martina, Maurizio and Stechele, Walter},
  title =	{{HW-Flow: A Multi-Abstraction Level HW-CNN Codesign Pruning Methodology}},
  journal =	{Leibniz Transactions on Embedded Systems},
  pages =	{03:1--03:30},
  ISSN =	{2199-2002},
  year =	{2022},
  volume =	{8},
  number =	{1},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LITES.8.1.3},
  doi =		{10.4230/LITES.8.1.3},
  annote =	{Keywords: Convolutional Neural Networks, Optimization, Hardware Modeling, Pruning}
}
Document
Short Paper
DOI: 10.4230/LIPIcs.COSIT.2022.25
Collaborative Wayfinding Under Distributed Spatial Knowledge (Short Paper)

Authors: Panagiotis Mavros, Saskia Kuliga, Ed Manley, Hilal Rohaidi Fitri, Michael Joos, and Christoph Hölscher

Published in: LIPIcs, Volume 240, 15th International Conference on Spatial Information Theory (COSIT 2022)

Abstract
In many everyday situations, two or more people navigate collaboratively but their spatial knowledge does not necessarily overlap. However, most research to date, has investigated social wayfinding under either 1-sided or fully shared spatial information. Here, we present the pilot experiment of a novel, computerised, non-verbal experimental paradigm to study collaborative wayfinding under the face of spatial information uncertainty. Participants (N=32) learned two different neighbourhoods individually, and then navigated together as dyads (D=16), from one neighbourhood to the other. Our pilot results reveal that overall participants share navigational control, but are in control more when the task leads them to a familiar destination. We discuss the effects of spatial ability and motivation to lead, as well as the outlook of the paradigm.
Cite as

Panagiotis Mavros, Saskia Kuliga, Ed Manley, Hilal Rohaidi Fitri, Michael Joos, and Christoph Hölscher. Collaborative Wayfinding Under Distributed Spatial Knowledge (Short Paper). In 15th International Conference on Spatial Information Theory (COSIT 2022). Leibniz International Proceedings in Informatics (LIPIcs), Volume 240, pp. 25:1-25:10, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2022)

Copy BibTex To Clipboard

@InProceedings{mavros_et_al:LIPIcs.COSIT.2022.25,
  author =	{Mavros, Panagiotis and Kuliga, Saskia and Manley, Ed and Fitri, Hilal Rohaidi and Joos, Michael and H\"{o}lscher, Christoph},
  title =	{{Collaborative Wayfinding Under Distributed Spatial Knowledge}},
  booktitle =	{15th International Conference on Spatial Information Theory (COSIT 2022)},
  pages =	{25:1--25:10},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-257-0},
  ISSN =	{1868-8969},
  year =	{2022},
  volume =	{240},
  editor =	{Ishikawa, Toru and Fabrikant, Sara Irina and Winter, Stephan},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.COSIT.2022.25},
  URN =		{urn:nbn:de:0030-drops-169105},
  doi =		{10.4230/LIPIcs.COSIT.2022.25},
  annote =	{Keywords: navigation, wayfinding, collaboration, dyad, online}
}
Document
Front Matter
DOI: 10.4230/DARTS.8.2.0
Front Matter - ECOOP 2022 Artifacts, Table of Contents, Preface, Artifact Evaluation Committee

Authors: Alessandra Gorla and Stefan Winter

Published in: DARTS, Volume 8, Issue 2, Special Issue of the 36th European Conference on Object-Oriented Programming (ECOOP 2022)

Abstract
Front Matter - ECOOP 2022 Artifacts, Table of Contents, Preface, Artifact Evaluation Committee
Cite as

Alessandra Gorla and Stefan Winter. Front Matter - ECOOP 2022 Artifacts, Table of Contents, Preface, Artifact Evaluation Committee. In Special Issue of the 36th European Conference on Object-Oriented Programming (ECOOP 2022). Dagstuhl Artifacts Series (DARTS), Volume 8, Issue 2, pp. 0:i-0:xii, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2022)

Copy BibTex To Clipboard

@Article{gorla_et_al:DARTS.8.2.0,
  author =	{Gorla, Alessandra and Winter, Stefan},
  title =	{{Front Matter - ECOOP 2022 Artifacts, Table of Contents, Preface, Artifact Evaluation Committee}},
  pages =	{0:i--0:xii},
  journal =	{Dagstuhl Artifacts Series},
  ISSN =	{2509-8195},
  year =	{2022},
  volume =	{8},
  number =	{2},
  editor =	{Gorla, Alessandra and Winter, Stefan},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/DARTS.8.2.0},
  URN =		{urn:nbn:de:0030-drops-161982},
  doi =		{10.4230/DARTS.8.2.0},
  annote =	{Keywords: Front Matter - ECOOP 2022 Artifacts, Table of Contents, Preface, Artifact Evaluation Committee}
}
Document
Short Paper
DOI: 10.4230/LIPIcs.GISCIENCE.2018.60
Abstract Data Types for Spatio-Temporal Remote Sensing Analysis (Short Paper)

Authors: Martin Sudmanns, Stefan Lang, Dirk Tiede, Christian Werner, Hannah Augustin, and Andrea Baraldi

Published in: LIPIcs, Volume 114, 10th International Conference on Geographic Information Science (GIScience 2018)

Abstract
Abstract data types are a helpful framework to formalise analyses and make them more transparent, reproducible and comprehensible. We are revisiting an approach based on the space, time and theme dimensions of remotely sensed data, and extending it with a more differentiated understanding of space-time representations. In contrast to existing approaches and implementations that consider only fixed spatial units (e.g. pixels), our approach allows investigations of the spatial units' spatio-temporal characteristics, such as the size and shape of their geometry, and their relationships. Five different abstract data types are identified to describe geographical phenomenon, either directly or in combination: coverage, time series, trajectory, composition and evolution.
Cite as

Martin Sudmanns, Stefan Lang, Dirk Tiede, Christian Werner, Hannah Augustin, and Andrea Baraldi. Abstract Data Types for Spatio-Temporal Remote Sensing Analysis (Short Paper). In 10th International Conference on Geographic Information Science (GIScience 2018). Leibniz International Proceedings in Informatics (LIPIcs), Volume 114, pp. 60:1-60:7, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2018)

Copy BibTex To Clipboard

@InProceedings{sudmanns_et_al:LIPIcs.GISCIENCE.2018.60,
  author =	{Sudmanns, Martin and Lang, Stefan and Tiede, Dirk and Werner, Christian and Augustin, Hannah and Baraldi, Andrea},
  title =	{{Abstract Data Types for Spatio-Temporal Remote Sensing Analysis}},
  booktitle =	{10th International Conference on Geographic Information Science (GIScience 2018)},
  pages =	{60:1--60:7},
  series =	{Leibniz International Proceedings in Informatics (LIPIcs)},
  ISBN =	{978-3-95977-083-5},
  ISSN =	{1868-8969},
  year =	{2018},
  volume =	{114},
  editor =	{Winter, Stephan and Griffin, Amy and Sester, Monika},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.GISCIENCE.2018.60},
  URN =		{urn:nbn:de:0030-drops-93881},
  doi =		{10.4230/LIPIcs.GISCIENCE.2018.60},
  annote =	{Keywords: Big Earth Data, Semantic Analysis, Data Cube}
}
Document
DOI: 10.4230/LITES-v004-i001-a004
How Is Your Satellite Doing? Battery Kinetics with Recharging and Uncertainty

Authors: Holger Hermanns, Jan Krčál, and Gilles Nies

Published in: LITES, Volume 4, Issue 1 (2017). Leibniz Transactions on Embedded Systems, Volume 4, Issue 1

Abstract
The kinetic battery model is a popular model of the dynamic behaviour of a conventional battery, useful to predict or optimize the time until battery depletion. The model however lacks certain obvious aspects of batteries in-the-wild, especially with respect to the effects of random influences and the behaviour when charging up to capacity limits.This paper considers the kinetic battery model with limited capacity in the context of piecewise constant yet random charging and discharging. We provide exact representations of the battery behaviour wherever possible, and otherwise develop safe approximations that bound the probability distribution of the battery state from above and below. The resulting model enables the time-dependent evaluation of the risk of battery depletion. This is demonstrated in an extensive dependability study of a nano satellite currently orbiting the earth.
Cite as

Holger Hermanns, Jan Krčál, and Gilles Nies. How Is Your Satellite Doing? Battery Kinetics with Recharging and Uncertainty. In LITES, Volume 4, Issue 1 (2017). Leibniz Transactions on Embedded Systems, Volume 4, Issue 1, pp. 04:1-04:28, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2017)

Copy BibTex To Clipboard

@Article{hermanns_et_al:LITES-v004-i001-a004,
  author =	{Hermanns, Holger and Kr\v{c}\'{a}l, Jan and Nies, Gilles},
  title =	{{How Is Your Satellite Doing? Battery Kinetics with Recharging and Uncertainty}},
  journal =	{Leibniz Transactions on Embedded Systems},
  pages =	{04:1--04:28},
  ISSN =	{2199-2002},
  year =	{2017},
  volume =	{4},
  number =	{1},
  publisher =	{Schloss Dagstuhl -- Leibniz-Zentrum f{\"u}r Informatik},
  address =	{Dagstuhl, Germany},
  URL =		{https://drops.dagstuhl.de/entities/document/10.4230/LITES-v004-i001-a004},
  doi =		{10.4230/LITES-v004-i001-a004},
  annote =	{Keywords: Battery Power, Depletion Risk, Bounded Charging and Discharging, Stochastic Load, Distribution Bounds}
}
  • Refine by Author
  • 1 Augustin, Hannah
  • 1 Baraldi, Andrea
  • 1 Camalleri, Manfredi
  • 1 Fasfous, Nael
  • 1 Fitri, Hilal Rohaidi
  • Show More...

  • Refine by Classification
  • 1 Applied computing → Psychology
  • 1 Computer systems organization → Reliability
  • 1 Computing methodologies → Artificial intelligence
  • 1 Computing methodologies → Neural networks
  • 1 General and reference → Empirical studies
  • Show More...

  • Refine by Keyword
  • 1 Artifact Evaluation Committee
  • 1 Battery Power
  • 1 Big Earth Data
  • 1 Bounded Charging and Discharging
  • 1 Convolutional Neural Networks
  • Show More...

  • Refine by Type
  • 6 document
  • 2 issue

  • Refine by Publication Year
  • 5 2022
  • 1 2017
  • 1 2018
  • 1 2023

Questions / Remarks / Feedback
X

Feedback for Dagstuhl Publishing


Thanks for your feedback!

Feedback submitted

Could not send message

Please try again later or send an E-mail
© 2023 Schloss Dagstuhl - LZI GmbH Imprint Privacy Contact