检索结果-内蒙古大学图书馆

2025 IEEE/CVF Winter Conference on Applications of computer Vision, WACV 2025

作者： Huber, Marco Damer, Naser Fraunhofer Institute for Computer Graphics Research Igd Germany Technische Universität Darmstadt Germany

ISBN: (纸本)9798331510831

The need for more transparent face recognition (FR), along with other visual-based decision-making systems has recently attracted more attention in research, society, and industry. The reasons why two face images are matched or not matched by a deep learning-based face recognition system are not obvious due to the high number of parameters and the complexity of the models. However, it is important for users, operators, and developers to ensure trust and accountability of the system and to analyze drawbacks such as biased behavior. While many previous works use spatial semantic maps to highlight the regions that have a significant influence on the decision of the face recognition system, frequency components which are also considered by CNNs, are neglected. In this work, we take a step forward and investigate explainable face recognition in the unexplored frequency domain. This makes this work the first to propose explainability of verification-based decisions in the frequency domain, thus explaining the relative influence of the frequency components of each input toward the obtained outcome. To achieve this, we manipulate face images in the spatial frequency domain and investigate the impact on verification outcomes. In extensive quantitative experiments, along with investigating two special scenarios cases, cross-resolution FR and morphing attacks (the latter in supplementary material), we observe the applicability of our proposed frequency-based explanations. © 2025 IEEE.

关键词： Frequency domain analysis

来源：评论

学校读者我要写书评

暂无评论

Utility Prediction Performance of Finger Image Quality Assessment Software 22

Utility Prediction Performance of Finger Image Quality Asses...

引用

22nd International Conference of the Biometrics Special Interest Group, BIOSIG 2023

作者： Henniger, Olaf Fraunhofer Institute for Computer Graphics Research IGD Darmstadt Germany

ISBN: (纸本)9798350336559

A biometric sample is the more utile for biometric recognition the greater the distance between the sample-specific non-mated and mated comparison score distributions. Finger image quality scores turn out to be only weakly correlated with the observed utility. This is worth investigating because finger image quality assessment software is widely used to predict the biometric utility of finger images in many public-sector applications. This paper shows that a weak correlation between predicted and observed utility does not matter if the quality scores are used to decide whether to discard or retain biometric samples for further processing. The important point is that useful samples are not mistakenly discarded or less useful samples are not mistakenly retained. This can be measured by quality-assessment false positive and false negative rates. In cost-benefit analyses, these metrics can be used to chose suitable quality-score thresholds for the use cases at hand. © 2023 IEEE.

关键词： Biometrics

来源：评论

学校读者我要写书评

暂无评论

Executing Ad-Hoc Queries on Large Geospatial Data Sets Without Acceleration Structures

引用

SN computer Science 2024年第5期5卷 647页

作者： Bormann, Pascal Krämer, Michel Würz, Hendrik M. Göhringer, Patrick Fraunhofer Institute for Computer Graphics Research IGD Darmstadt 64283 Germany Technical University of Darmstadt Darmstadt 64289 Germany

In this case study, we investigate if it is possible to harness the capabilities of modern commodity hardware to perform ad-hoc queries on large raw geospatial data sets. Normally, this requires building an index structure, which is a time-consuming process. We aim to provide means to individual users who receive a new or updated geospatial data set and want to directly start working with it without having to build such an index structure first. To this end, we conduct various experiments on two distinct types of data: 3D building models and point clouds. For the former, we demonstrate that well-known algorithms such as fast string search allow a wide range of queries to be answered in at most a few seconds on data sets with over a million buildings. The usage of progressive indexing additionally improves query run time by more than a factor of two. Regarding point clouds, we achieve similar run times using the popular LAS file format and a query throughput of up to a billion points per second when using a columnar memory layout. The run time of ad-hoc queries is often on par with that of database-driven solutions, sometimes even outperforming them. Considering that ad-hoc queries require no preprocessing, our results show that they are a viable alternative to acceleration structures when working with geospatial data. © The Author(s) 2024.

关键词： Building models Geospatial data Information retrieval Point clouds Searching

来源：评论

学校读者我要写书评

暂无评论

Brain computer Interfacing with a Virtual Environment 32

Brain Computer Interfacing with a Virtual Environment

引用

32nd International Conference in Central Europe on computer graphics, Visualization and computer Vision, WSCG 2024

作者： Gamillscheg, Florian Ruprecht, Irena Settgast, Volker Pietroszek, Krzysztof Augsdörfer, Ursula Institute of Computer Graphics and Knowledge Visualisation Graz University of Technology Austria Fraunhofer Austria Graz Austria American University WashingtonDC United States Graz University of Technology Austria

Virtual Reality (VR) applications constantly strive for more realism, immersion and intuitive user experiences. Traditional VR controllers can hinder full immersion, since they form an additional barrier between the user’s thoughts or intentions and the virtual world. Brain computer interfaces (BCIs) have the potential to close this gap by enabling an immediate translation of human thoughts to commands that can be processed by a computer. This paper investigates the feasibility of employing an affordable commercial BCI device for VR interaction. In a preliminary study conducted in a Cave Automatic Virtual Environment (CAVE), we evaluate both the effectiveness and limitations of the popular BCI device Emotiv Insight. © 2024 University of West Bohemia. All rights reserved.

关键词： Virtual environments

来源：评论

学校读者我要写书评

暂无评论

FAST GDRNPP: Improving the Speed of State-of-the-Art 6D Object Pose Estimation

FAST GDRNPP: Improving the Speed of State-of-the-Art 6D Obje...

引用

2024 Eurographics Italian Chapter Conference on Smart Tools and Applications in graphics, STAG 2024

作者： Pollabauer, T. Pramod, A. Knauthe, V. Wahl, M. Fraunhofer Institute for Computer Graphics Research Germany Technical University Darmstadt Germany University of Siegen Germany

ISBN: (纸本)9783038682653

6D object pose estimation involves determining the three-dimensional translation and rotation of an object within a scene and relative to a chosen coordinate system. This problem is of particular interest for many practical applications in industrial tasks such as quality control, bin picking, and robotic manipulation, where both speed and accuracy are critical for real-world deployment. Current models, both classical and deep-learning-based, often struggle with the trade-off between accuracy and latency. Our research focuses on enhancing the speed of a prominent state-of-the-art deep learning model, GDRNPP, while keeping its high accuracy. We employ several techniques to reduce the model size and improve inference time. These techniques include using smaller and quicker backbones, pruning unnecessary parameters, and distillation to transfer knowledge from a large, high-performing model to a smaller, more efficient student model. Our findings demonstrate that the proposed configuration maintains accuracy comparable to the state-of-the-art while significantly improving inference time. This advancement could lead to more efficient and practical applications in various industrial scenarios, thereby enhancing the overall applicability of 6D Object Pose Estimation models in real-world settings. © 2024 The Authors.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Towards a Face Verification System-on-Card with Compound-Eye Camera 9th

Towards a Face Verification System-on-Card with Compound-Eye...

引用

9th Future Technologies Conference, FTC 2024

作者： Henniger, Olaf Strübig, Christoph Tilman Kuijper, Arjan Fraunhofer Institute for Computer Graphics Research IGD Darmstadt Germany Technical University of Darmstadt Darmstadt Germany

ISBN: (纸本)9783031731211

Electronic passports and other identity documents store biometric reference data of the document holder (such as a face image and fingerprint images) in a smart-card chip with contactless interface to enable the automatic verification that the holder is the legitimate holder. Off-card biometric comparison involves sending sensitive biometric reference data via near-field communication to a background system. Even if encrypted, the transfer of biometric data requires considerable trust in the background system and can threaten the privacy of the cardholder. A biometric system-on-card integrating the biometric sensor and signal processing into the card enhances privacy because sensitive biometric data never leave the card. System-on-card technology with capacitive fingerprint sensors integrated into a smart card is already mature. Optical camera modules for capturing face images, however, have not been integrated into smart cards yet. Ultra-thin camera modules inspired by the compound eyes of insects could be integrated into smart cards, but would the computing resources on a smart card be sufficient for face recognition? This paper assesses the feasibility of a face-verification system-on-card using an ultra-thin compound-eye camera and a pre-trained lightweight neural network. Taking the resource limitations of smart cards into account, a face verification pipeline from capture to verification decision is outlined, and its accuracy is estimated. This paper shows that lightweight face verification using a compound-eye camera is feasible with acceptable accuracy but requires more memory and computing power than available on today’s off-the-shelf smart cards. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： System-on-chip

来源：评论

学校读者我要写书评

暂无评论

Detection and Pose Estimation of Flat, Texture-Less Industry Objects on HoloLens Using Synthetic Training 23rd

Detection and Pose Estimation of Flat, Texture-Less Indust...

引用

22nd Scandinavian Conference on Image Analysis, SCIA 2023

作者： Pöllabauer, Thomas Rücker, Fabian Franek, Andreas Gorschlüter, Felix Fraunhofer Institute for Computer Graphics Research IGD Darmstadt Germany Interactive Graphics Research Group TU Darmstadt Darmstadt Germany

ISBN: (纸本)9783031314377

Current state-of-the-art 6d pose estimation is too compute intensive to be deployed on edge devices, such as Microsoft HoloLens (2) or Apple iPad, both used for an increasing number of augmented reality applications. The quality of AR is greatly dependent on its capabilities to detect and overlay geometry within the scene. We propose a synthetically trained client-server-based augmented reality application, demonstrating state-of-the-art object pose estimation of metallic and texture-less industry objects on edge devices. Synthetic data enables training without real photographs, i.e. for yet-to-be-manufactured objects. Our qualitative evaluation on an AR-assisted sorting task, and quantitative evaluation on both renderings, as well as real-world data recorded on HoloLens 2, sheds light on its real-world applicability. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Textures

来源：评论

学校读者我要写书评

暂无评论

E-Sports Player Performance Metrics for Predicting the Outcome of League of Legends Matches Considering Player Roles

引用

SN computer Science 2023年第3期4卷 238页

作者： Bahrololloomi, Farnod Klonowski, Fabio Sauer, Sebastian Horst, Robin Dörner, Ralf RheinMain University of Applied Sciences Wiesbaden Germany Fraunhofer Institute for Computer Graphics Research IGD Darmstadt Germany

Predicting results in electronic sports (e-sports) matches is not an easy task. Different methods can be used for this purpose. A well-known video game in the field of Multiplayer Online Battle Arena (MOBA) is the game League of Legends (LoL), which has a relevant professional scene. An important part of professional gaming is analyzing past matches overall and an individual player’s performance to prepare for future matches. In this paper, we follow a design-oriented research methodology (analysis, design, and evaluation) and propose performance metrics that use data from past matches to evaluate a player’s performance. We analyze the necessary data which we acquire by selecting a player, analyzing the player’s latest games, and repeating the process recursively with the players found in his latest games. The data is utilized within a Machine Learning (ML) Model that computes an overall score from individual player variables. From this, we designed a heuristic approach and evaluated it by applying it to the challenge of winning predictions in e-sports. The difference in the influence of the individual player roles on the outcome of the game was also investigated. It was found that this difference is negligible and that the heuristic performance metric can predict the outcome of a game with an accuracy of 86%. Furthermore, the concept of a match calculator is explored, which calculates the outcome of a match using the ML model and different player stats. © 2023, The Author(s).

关键词： computer games Data science Electronic sports (e-sports) League of Legends Machine Learning Performance analysis Player modeling Winning prediction

来源：评论

学校读者我要写书评

暂无评论

One-to-many Reconstruction of 3D Geometry of cultural Artifacts using a synthetically trained Generative Model

One-to-many Reconstruction of 3D Geometry of cultural Artifa...

引用

2023 Eurographics Workshop on graphics and Cultural Heritage, GCH 2023

作者： Pöllabauer, T. Kühn, J. Li, J. Kuijper, A. Fraunhofer Institute for Computer Graphics Research IGD Germany Interactive Graphics Research Group TU Darmstadt Germany

ISBN: (纸本)9783038682172

Estimating the 3D shape of an object using a single image is a difficult problem. Modern approaches achieve good results for general objects, based on real photographs, but worse results on less expressive representations such as historic sketches. Our automated approach generates a variety of detailed 3D representation from a single sketch, depicting a medieval statue, and can be guided by multi-modal inputs, such as text prompts. It relies solely on synthetic data for training, making it adoptable even in cases of only small numbers of training examples. Our solution allows domain experts such as a curators to interactively reconstruct potential appearances of lost artifacts. © ISVR,2023 *** rights reserved

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Transparency Distortion Robustness for SOTA Image Segmentation Tasks 4th

Transparency Distortion Robustness for SOTA Image Segmentat...

引用

4th International Conference on Pattern Recognition and Artificial Intelligence, ICPRAI 2024

作者： Knauthe, Volker Rak, Arne Wirth, Tristan Pöllabauer, Thomas Metzler, Simon Kuijper, Arjan Fellner, Dieter W. Technical University of Darmstadt Darmstadt Germany Fraunhofer Institute for Computer Graphics Research IGD Darmstadt Germany CGV Institute Graz University of Technology Graz Austria

ISBN: (纸本)9789819787043

Semantic Image Segmentation facilitates a multitude of real-world applications ranging from autonomous driving over industrial process supervision to vision aids for human beings. These models are usually trained in a supervised fashion using example inputs. Distribution Shifts between these examples and the inputs in operation may cause erroneous segmentations. The robustness of semantic segmentation models against distribution shifts caused by differing camera or lighting setups, lens distortions, adversarial inputs and image corruptions has been topic of recent research. However, robustness against spatially varying radial distortion effects that can be caused by uneven glass structures (e.g. windows) or the chaotic refraction in heated air has not been addressed by the research community yet. We propose a method to synthetically augment existing datasets with spatially varying distortions. Our experiments show, that these distortion effects degrade the performance of state-of-the-art segmentation models. Pretraining and enlarged model capacities proof to be suitable strategies for mitigating performance degradation to some degree, while fine-tuning on distorted images only leads to marginal performance improvements. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：