检索结果-内蒙古大学图书馆

SynFacePAD 2023: Competition on Face Presentation Attack Detection Based on Privacy-aware Synthetic Training Data

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Fang, Meiling Huber, Marco Fierrez, Julian Ramachandra, Raghavendra Damer, Naser Alkhaddour, Alhasan Kasantcev, Maksim Pryadchenko, Vasiliy Yang, Ziyuan Huangfu, Huijie Chen, Yingyu Zhang, Yi Pan, Yuchen Jiang, Junjun Liu, Xianming Sun, Xianyun Wang, Caiyong Liu, Xingyu Chang, Zhaohua Zhao, Guangzhe Tapia, Juan Gonzalez-Soler, Lazaro Aravena, Carlos Schulz, Daniel Fraunhofer Institute for Computer Graphics Research IGD Darmstadt Germany Department of Computer Science TU Darmstadt Darmstadt Germany Biometrics and Data Pattern Analytics Lab Universidad Autonoma de Madrid Spain Norway ID R&D Inc. New York United States School of Cyber Science and Engineering Sichuan University Chengdu China School of Computer Science and Technology Harbin Institute of Technology Harbin China School of Electrical and Information Engineering Beijing University of Civil Engineering and Architecture China Biometrics and Security Research Group Hochschule Darmstadt Darmstadt Germany I+D Vision Center Santiago Chile

This paper presents a summary of the Competition on Face Presentation Attack Detection Based on Privacy-aware Synthetic Training Data (SynFacePAD 2023) held at the 2023 International Joint Conference on Biometrics (IJCB 2023). The competition attracted a total of 8 participating teams with valid submissions from academia and industry. The competition aimed to motivate and attract solutions that target detecting face presentation attacks while considering synthetic-based training data motivated by privacy, legal and ethical concerns associated with personal data. To achieve that, the training data used by the participants was limited to synthetic data provided by the organizers. The submitted solutions presented innovations and novel approaches that led to outperforming the considered baseline in the investigated benchmarks. © 2023, CC BY-NC-SA.

关键词： Competition

Predicting Interacting Green’s Functions with Neural Networks

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Agapov, Egor Bertomeu, Oriol Carballo, Andrés Mendl, Christian B. Sander, Aaron Ludwig-Maximilians-Universität München Munich Germany School of Natural Sciences Technical University of Munich Garching Germany Department of Computer Science Technical University of Munich Garching Germany Institute for Advanced Study Technical University of Munich Garching Germany Chair for Design Automation Technical University of Munich Munich Germany

Strongly correlated materials exhibit complex electronic phenomena that are challenging to capture with traditional theoretical methods, yet understanding these systems is crucial for discovering new quantum materials. Addressing the computational bottlenecks in studying such systems, we present a proof-of-concept machine learning-based approach to accelerate Dynamical Mean Field Theory (DMFT) calculations. Our method predicts interacting Green’s functions on arbitrary two-dimensional lattices using a two-step ML framework. First, an autoencoder-based network learns and generates physically plausible band structures of materials, providing diverse training data. Next, a dense neural network predicts interacting Green’s functions of these physically-possible band structures, expressed in the basis of Legendre polynomials. We demonstrate that this architecture can serve as a substitute for the computationally demanding quantum impurity solver in DMFT, significantly reducing computational cost while maintaining accuracy. This approach offers a scalable pathway to accelerate simulations of strongly correlated systems and lays the groundwork for future extensions to multi-band systems. Copyright © 2024, The Authors. All rights reserved.

关键词： Nanocrystals

Digital few-mode fiber multiplexer using multiplane light conversion

学校读者我要写书评

暂无评论

Digital few-mode fiber multiplexer using multiplane light co...

IEEE Photonics Conference (IPC)

作者： Dennis Pohle Fabio A. Barbosa Filipe M. Ferreira Stefan Rothe Juergen W Czarske Chair of Measurement and Sensor System Technique Faculty of Electrical and Computer Engineering TU Dresden Dresden Germany Department of Electronic and Electrical Engineering Optical Networks Group University College London United Kingdom Institute of Applied Physics School of Science TU Dresden Dresden Germany

We propose an all-optical data-driven technique for space division multiplexing in few-mode fibers. A digital twin was realized by multiplane light conversion and neural networks. It is promising for a digitally progr...

关键词：

Spatial imagination with semantic cognition for mobile robots

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Shen, Zhengcheng Kästner, Linh Lambrecht, Jens The Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

The imagination of the surrounding environment based on experience and semantic cognition has great potential to extend the limited observations and provide more information for mapping, collision avoidance, and path planning. This paper provides a training-based algorithm for mobile robots to perform spatial imagination based on semantic cognition and evaluates the proposed method for the mapping task. We utilize a photo-realistic simulation environment, Habitat, for training and evaluation. The trained model is composed of Resent-18 as encoder and Unet as the backbone. We demonstrate that the algorithm can perform imagination for unseen parts of the object universally, by recalling the images and experience and compare our approach with traditional semantic mapping methods. It is found that our approach will improve the efficiency and accuracy of semantic mapping. © 2021, CC BY-NC-ND.

关键词： Mobile robots

Residue super-resolution convolutional neural network based complexity reduction for H.266 VVC intra-coding

学校读者我要写书评

暂无评论

ICT Express 2025年第3期11卷 460-466页

作者： Dhanalakshmi, A. Balaji, L. Raja, C. Giri, Jayant Alrashoud, Mubarak Department of ECE Panimalar Engineering College Chennai India Department of ECE Vel Tech Rangarajan Dr.Sagunthala R&D Institute of Science and Technology Chennai India Department of Mechanical Engineering Yeshwantrao Chavan College of Engineering Nagpur India Division of Research and Development Lovely Professional University Phagwara India Chair of Smart Technologies College of Computer and Information Sciences King Saud University Riyadh Saudi Arabia

Versatile Video Coding (VVC) promised to provide the same video quality as HEVC with 50 % bitrate reduction, which was introduced in 2020. Our suggested method for VVC Intra-coding is residue super-resolution convolutional neural network (RSR-CNN) utilizing downsampling and upsampling procedures. We present an effective complexity reduced VVC intra-coding scheme based on residue SR-CNN. Reducing an original video's resolution in both the vertical and horizontal directions is all that is required to execute down sampling. Increasing the video dimensions for improved visual quality, convolutional neural networks are utilized in the upsampling process to create residue super-resolution. Specifically, for every block, we train a CNN model to perform residue SR after downsampling and compressing the residue at low resolution, and then we carry out motion estimation (ME) and motion compensation (MC) to extract the residue. Using the MC prediction signal, a new residue SR-CNN is designed. Additionally, this work comprehensively examines the complexity and performance of VVC intra-coding tools and integrates them with the residue SR-CNN method. The experiments demonstrate a substantial time savings of 40 % in encoding with BDBR coding gains of 4.2 %, and 2.9 % in AI and RA configurations respectively. © 2025 The Author(s)

关键词： Intra-coding Low complexity PSNR Residue super-resolution convolutional neural network VVC

Connecting deep-reinforcement-learning-based obstacle avoidance with conventional global planners using waypoint generators

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Kästner, Linh Buiyan, Teham Zhao, Xinlin Shen, Zhengcheng Lambrecht, Jens Marx, Cornelius Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

Deep Reinforcement Learning has emerged as an efficient dynamic obstacle avoidance method in highly dynamic environments. It has the potential to replace overly conservative or inefficient navigation approaches. However, the integration of Deep Reinforcement Learning into existing navigation systems is still an open frontier due to the myopic nature of Deep-Reinforcement-Learning-based navigation, which hinders its widespread integration into current navigation systems. In this paper, we propose the concept of an intermediate planner to interconnect novel Deep-Reinforcement-Learning-based obstacle avoidance with conventional global planning methods using waypoint generation. Therefore, we integrate different waypoint generators into existing navigation systems and compare the joint system against traditional ones. We found an increased performance in terms of safety, efficiency and path smoothness especially in highly dynamic environments. © 2021, CC BY-NC-ND.

关键词： Deep learning

LLMs and the Abstraction and Reasoning Corpus: Successes, Failures, and the Importance of Object-based Representations

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Xu, Yudong Li, Wenhao Vaezipoor, Pashootan Sanner, Scott Khalil, Elias B. Department of Mechanical & Industrial Engineering University of Toronto Canada Department of Computer Science University of Toronto Vector Institute for Artificial Intelligence Canada Department of Mechanical & Industrial Engineering University of Toronto Scale AI Research Chair in Data-Driven Algorithms for Modern Supply Chains Canada

Can a Large Language Model (LLM) solve simple abstract reasoning problems? We explore this broad question through a systematic analysis of GPT on the Abstraction and Reasoning Corpus (ARC) (Chollet, 2019), a representative benchmark of abstract reasoning ability from limited examples in which solutions require some "core knowledge" of concepts such as objects, goal states, counting, and basic geometry. GPT-4 solves only 13/50 of the most straightforward ARC tasks when using textual encodings for their two-dimensional input-output grids. Our failure analysis reveals that GPT-4’s capacity to identify objects and reason about them is significantly influenced by the sequential nature of the text that represents an object within a text encoding of a task. To test this hypothesis, we design a new benchmark, the 1D-ARC, which consists of one-dimensional (array-like) tasks that are more conducive to GPT-based reasoning, and where it indeed performs better than on the (2D) ARC. To alleviate this issue, we propose an object-based representation that is obtained through an external tool, resulting in nearly doubling the performance on solved ARC tasks and near-perfect scores on the easier 1D-ARC. Although the state-of-the-art GPT-4 is unable to "reason" perfectly within non-language domains such as the 1D-ARC or a simple ARC subset, our study reveals that the use of object-based representations can significantly improve its reasoning ability. Visualizations, GPT logs, and data are available at https://***/LLM4ARC. © 2023, CC BY.

关键词： Abstracting

Integrative Object and Pose to Task Detection for an Augmented-Reality-based Human Assistance System using Neural Networks

学校读者我要写书评

暂无评论

Integrative Object and Pose to Task Detection for an Augment...

International Conference on Communications and Electronics, ICCE

作者： Linh Kästner Leon Eversberg Marina Mursa Jens Lambrecht Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

ISBN: (数字)9781728154718

ISBN: (纸本)9781728154725

As a result of an increasingly automatized and digitized industry, processes are becoming more complex. Augmented Reality has shown considerable potential in assisting workers with complex tasks by enhancing user understanding and experience with spatial information. However, the acceptance and integration of AR into industrial processes is still limited due to the lack of established methods and tedious integration efforts. Meanwhile, deep neural networks have achieved remarkable results in computer vision tasks and bear great prospects to enrich Augmented Reality applications. In this paper, we propose an Augmented-Reality-based human assistance system to assist workers in complex manual tasks where we incorporate deep neural networks for computer vision tasks. More specifically, we combine Augmented Reality with object and action detectors to make workflows more intuitive and flexible. To evaluate our system in terms of user acceptance and efficiency, we conducted several user studies. We found a significant reduction in time to task completion in untrained workers and a decrease in error rate. Furthermore, we investigated the users learning curve with our assistance system.

关键词： computer vision Error analysis Neural networks Manuals computer architecture Task analysis Augmented reality

Autonomous Navigation in Complex Environments using Memory-Aided Deep Reinforcement Learning

学校读者我要写书评

暂无评论

Autonomous Navigation in Complex Environments using Memory-A...

IEEE/SICE International Symposium on System Integration

作者： Linh Kästner Zhengcheng Shen Cornelius Marx Jens Lambrecht Chair Industry Grade Networks and Clouds Faculty of Electrical Engineering and Computer Science Berlin Institute of Technology Berlin Germany

ISBN: (数字)9781728176581

ISBN: (纸本)9781728176598

Mobile robots have gained increased importance within industrial tasks such as commissioning, delivery or operation in hazardous environments. The ability to navigate in unknown and complex environments is paramount in industrial robotics. Reinforcement learning approaches have shown remarkable success in dealing with unknown situations and react accordingly without manually engineered guidelines and overconservative measures. However, these approaches are often restricted to short range navigation and are prone to local minima due to a lack of a memory module. Thus, the navigation in complex environments such as mazes, long corridors or concave areas is still an open frontier. In this paper, we incorporate a variety of recurrent neural networks to cope with these challenges. We train a reinforcement learning based agent within a 2D simulation environment of our previous work and extend it with a memory module. The agent is able to navigate solely on sensor data observations which are directly mapped to actions. We evaluate the performance on different complex environments and achieve enhanced results within complex environments compared to memory-free baseline approaches.

关键词： Training Navigation Two dimensional displays Reinforcement learning Memory modules System integration Task analysis