检索结果-内蒙古大学图书馆

Neural networks for semantic gaze analysis in XR settings

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Stubbemann, Lena Dürrschnabel, Dominik Refflinghaus, Robert Department of Quality and Process Management University of Kassel Germany Knowledge & Data Engineering Group University of Kassel Germany Interdisciplinary Research Center for Information System Design University of Kassel Germany

Virtual-reality (VR) and augmented-reality (AR) technology is increasingly combined with eye-tracking. This combination broadens both fields and opens up new areas of application, in which visual perception and related cognitive processes can be studied in interactive but still well controlled settings. However, performing a semantic gaze analysis of eye-tracking data from interactive three-dimensional scenes is a resource-intense task, which so far has been an obstacle to economic use. In this paper we present a novel approach which minimizes time and information necessary to annotate volumes of interest (VOIs) by using techniques from object recognition. To do so, we train convolutional neural networks (CNNs) on synthetic data sets derived from virtual models using image augmentation techniques. We evaluate our method in real and virtual environments, showing that the method can compete with state-of-the-art approaches, while not relying on additional markers or preexisting databases but instead offering cross-platform *** Codes 68T45 Copyright © 2021, The Authors. All rights reserved.

关键词： Semantics

A Deep Neural Network based Multimodal Video Recognition System for Caring

学校读者我要写书评

暂无评论

A Deep Neural Network based Multimodal Video Recognition Sys...

IEEE International Conference on Human-Machine Systems

作者： Chao Yan Jiahua Xu Bastian Klopfer Andreas Nürnberger Data & Knowledge Engineering Group Faculty of Computer Science Otto von Guericke University Magdeburg Germany

ISBN: (数字)9781728158716

ISBN: (纸本)9781728158723

Home caring usually refers to taking care of the elder, the young kids and the patients at home, and this depends on both caregivers and caring systems. The system which can provide sufficient and accurate information to doctors or caregivers without a delay will bring benefits for all the persons who need to be cared and allow doctor or care givers to take right action based on the information immediately. Available corresponding products in the market mainly are some smart home devices or some medical facilities based on electroencephalo-graph (EEG), electrocardiograph (ECG) or blood pressure check, however, these parameters cannot give doctors, caregivers or family members a direct feedback. To address the problem, this paper introduces a deep learning based design - a visual recognition system developed for clinical monitoring which can supervise both the emotions and gestures of patients at the same time and give responsible persons instant and direct feedback so that the right treatment will be taken by them. This product uses a Raspberry Pi computer with its camera as hardware and implementing several deep learning models to fulfill three main functions which are: Facial Recognition, Emotion Detection and Pose Estimation onto a portable device, which enhances the utilization of the application and this brings more possibilities. Compared to theavailable products, this application emphasizes monitoring per-sons via visual analysis and gives more direct feedback all thetime instead of traditional ways which are not always timely or practical.

关键词： Faces Computational modeling Machine learning Face recognition Pose estimation Cameras Medical services

Neurorehabilitation System in Virtual Reality with Low-Cost BCI Devices

学校读者我要写书评

暂无评论

Neurorehabilitation System in Virtual Reality with Low-Cost ...

IEEE International Conference on Human-Machine Systems

作者： Jiahua Xu Tung-Lung Liu Zhenglei Wu Zheng Wu Yang Li Andreas Nürnberger Data & Knowledge Engineering Group Faculty of Computer Science Otto-von-Guericke University Magdeburg Germany

ISBN: (数字)9781728158716

ISBN: (纸本)9781728158723

Brain-computer interface (BCI) has gained great attention from the public due to its ability and portability for external devices control or neuron rehabilitation using EEG from scalps. virtual reality (VR) provides users an immersive interaction environment via human-machine interfaces (HMI).This paper proposes an motor imagery based real-time control system for rehabilitation by using low-cost BCI to control a robotic arm in VR.

关键词： Electroencephalography Real-time systems Manipulators Brain-computer interfaces Training Task analysis

Sub-word Based End-to-End Speech Recognition for an Under-Resourced Language: Amharic

学校读者我要写书评

暂无评论

Sub-word Based End-to-End Speech Recognition for an Under-Re...

IEEE International Conference on Systems, Man and Cybernetics

作者： Nirayo Hailu Gebreegziabher Andreas Nürnberger Faculty of Computer Science Data and Knowledge Engineering Group Otto-von-Guericke University of Magdeburg Magdeburg Germany

ISBN: (数字)9781728185262

ISBN: (纸本)9781728185279

In this work, we focused on end-to-end speech recognition for less-resourced language, Amharic. The result can be integrated with other tasks such as spoken content retrieval. We explored three models, which consist of Convolutional Neural Networks, Recurrent Neural Networks, and Connectionist Temporal Classification, towards end-to-end speech recognition on less-resourced language. Further, we studied the possibility of having an end-to-end system with 1-best output keeping the network parameters and computational resource minimal. The paper gives attention to finding a more suitable sub-lexical unit for the Amharic end-to-end speech recognition system which can be used as an audio indexing unit. We present the first result comparing grapheme, phoneme, and syllable-based end-to-end speech recognition systems for our target language. The models are evaluated on approximately 52 hours of Amharic speech corpus containing read-speech, audiobooks, and multi-genre radio programs. On the test set, we report a character error rate (CER) of 19.21% and a syllable error rate (SER) of 39.98% for a syllable-based end-to-end model without lexicons and language model integrated.

关键词： Training Error analysis Computational modeling Speech recognition Computer architecture data models Task analysis

A Light-weight Convolutional Neural Network based Speech Recognition for Spoken Content Retrieval Task

学校读者我要写书评

暂无评论

A Light-weight Convolutional Neural Network based Speech Rec...

IEEE International Conference on Systems, Man and Cybernetics

作者： Nirayo Hailu Gebreegziabher Andreas Nürnberger Faculty of Computer Science Data and Knowledge Engineering Group Otto-von-Guericke University of Magdeburg Magdeburg Germany

ISBN: (数字)9781728185262

ISBN: (纸本)9781728185279

Convolutional Neural Network has shown to achieve a state of the art performance in computer vision. They have also progressively become popular in speech recognition and other natural language processing tasks. In this study, we aim at designing a light-weight Convolutional Neural Network architecture for the under-resourced end-to-end speech recognition task. We present a carefully designed 1-dimensional Convolutional deep neural network architecture that could achieve reasonable accuracy to be cascaded with spoken content retrieval systems. We explored the usage of Convolutional Neural Networks with Connectionist Temporal Classification under resource-constrained conditions. The possibility of having an end-to-end system with the best decoding result keeping the network parameters and computational time minimum is also shown. The paper presents the results on the Amharic syllable-based end-to-end speech recognition system implementing the designed model. The architecture is trained and evaluated on ≈70 hours of Amharic read-speech, audiobooks, and multi-genre radio programs. On the development set, we report a character error rate of 12.60% and a syllable error rate of 27.28% without language-models integrated. Likewise, on the test set 18.38% character error rate and 27.71% syllable error rate is reached.

关键词： Training Error analysis Computational modeling Speech recognition Computer architecture Convolutional neural networks Task analysis

Newsalyze: Enabling news consumers to understand media bias 20

学校读者我要写书评

暂无评论

Newsalyze: Enabling news consumers to understand media bias

2020 ACM/IEEE-CS Joint Conference on Digital Libraries, JCDL 2020

作者： Hamborg, Felix Zhukova, Anastasia Donnay, Karsten Gipp, Bela Dept. of Computer and Information Science University of Konstanz Germany Data and Knowledge Engineering Group University of Wuppertal Wuppertal Germany Dept. of Political Science University of Zurich Switzerland Dept. of Politics and Public Administration University of Konstanz Germany

ISBN: (纸本)9781450375856

News is a central source of information for individuals to inform themselves on current topics. Knowing a news article's slant and authenticity is of crucial importance in times of fake news, news bots, and centralization of media ownership. We introduce Newsalyze, a bias-aware news reader focusing on a subtle, yet powerful form of media bias, named bias by word choice and labeling (WCL). WCL bias can alter the assessment of entities reported in the news, e.g., freedom fighters vs. terrorists. At the core of the analysis is a neural model that uses a news-adapted BERT language model to determine target-dependent sentiment, a high-level effect of WCL bias. While the analysis currently focuses on only this form of bias, the visualizations already reveal patterns of bias when contrasting articles (overview) and in-text instances of bias (article view). © 2020. ACM ISBN.

关键词： news bias annotation newsroom frame analysis news articles news consumption content analysis media bias

Learning Gradient Boosted Multi-label Classification Rules

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Rapp, Michael Mencía, Eneldo Loza Fürnkranz, Johannes Nguyen, Vu-Linh Hüllermeier, Eyke Knowledge Engineering Group TU Darmstadt Germany Computational Data Analysis Group JKU Linz Austria Heinz Nixdorf Institute Paderborn University Germany

In multi-label classification, where the evaluation of predictions is less straightforward than in single-label classification, various meaningful, though different, loss functions have been proposed. Ideally, the learning algorithm should be customizable towards a specific choice of the performance measure. Modern implementations of boosting, most prominently gradient boosted decision trees, appear to be appealing from this point of view. However, they are mostly limited to single-label classification, and hence not amenable to multi-label losses unless these are label-wise decomposable. In this work, we develop a generalization of the gradient boosting framework to multi-output problems and propose an algorithm for learning multi-label classification rules that is able to minimize decomposable as well as non-decomposable loss functions. Using the well-known Hamming loss and subset 0/1 loss as representatives, we analyze the abilities and limitations of our approach on synthetic data and evaluate its predictive performance on multi-label benchmarks. Copyright © 2020, The Authors. All rights reserved.

关键词： Classification (of information)

Weakly-supervised segmentation using inherently-explainable classification models and their application to brain tumour classification

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Chatterjee, Soumick Yassin, Hadya Dubost, Florian Nürnberger, Andreas Speck, Oliver Data and Knowledge Engineering Group Faculty of Computer Science Otto von Guericke University Magdeburg Germany Biomedical Magnetic Resonance Faculty of Nature Sciences Otto von Guericke University Magdeburg Germany Genomics Research Centre Human Technopole Milan Italy Institute for Medical Engineering Faculty of Electrical Engineering and Information Technology Otto von Guericke University Magdeburg Germany Department of Biomedical Data Science Stanford University StanfordCA United States Centre for Behavioural Brain Sciences Magdeburg Germany German Centre for Neurodegenerative Disease Magdeburg Germany

Deep learning models have shown their potential in recent years for several applications. However, most of the models are opaque in nature and difficult to trust due to their complex reasoning - commonly known as the black-box problem. Some fields, such as medicine, require a high degree of transparency to accept and adopt such technologies. Consequently, there is a movement towards creating explainable/interpretable models or applying post-hoc explainability/interpretability methods on classifiers to build trust in deep learning models. Moreover, deep learning methods can be used for segmentation tasks, which typically require hard-to-obtain, time-consuming manually-annotated segmentation labels for training. This paper introduces three inherently-explainable classifiers to tackle both of these problems as one. The localisation heatmaps provided by the networks - representing the models' focus areas and being used in classification decision-making - can be directly interpreted, without requiring any post-hoc methods to derive information for model explanation. The models are trained by using the input image and only the classification labels as groundtruth in a supervised fashion - without using any information about the location of the region of interest (i.e. the segmentation labels), making the segmentation training of the models weakly-supervised through classification labels. The final segmentation of the regions of interest is obtained by thresholding these heatmaps. The models were employed for the task of multi-class brain tumour classification using two different datasets, resulting in the best F1-score of 0.93 for the supervised classification task while securing a median Dice score of 0.67 ± 0.08 for the weakly-supervised segmentation task. Furthermore, the obtained accuracy on a subset of tumour-only images outperformed the state-of-the-art glioma tumour grading binary classifiers with the best model achieving 98.7% accuracy. Copyright © 2022, The Autho

关键词： Magnetic resonance imaging

Robust and Fast Measure of Information via Low-rank Representation

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Dong, Yuxin Gong, Tieliang Yu, Shujian Chen, Hong Li, Chen School of Computer Science and Technology Xi’an Jiaotong University Xi’an710049 China Shaanxi Provincial Key Laboratory of Big Data Knowledge Engineering Ministry of Education Xi’an710049 China Machine Learning Group UiT - The Arctic University of Norway Norway College of Science Huazhong Agriculture University Wuhan430070 China Engineering Research Center of Intelligent Technology for Agriculture Ministry of Education Wuhan430070 China

The matrix-based Rényi’s entropy allows us to directly quantify information measures from given data, without explicit estimation of the underlying probability distribution. This intriguing property makes it widely applied in statistical inference and machine learning tasks. However, this information theoretical quantity is not robust against noise in the data, and is computationally prohibitive in large-scale applications. To address these issues, we propose a novel measure of information, termed low-rank matrix-based Rényi’s entropy, based on low-rank representations of infinitely divisible kernel matrices. The proposed entropy functional inherits the specialty of of the original definition to directly quantify information from data, but enjoys additional advantages including robustness and effective calculation. Specifically, our low-rank variant is more sensitive to informative perturbations induced by changes in underlying distributions, while being insensitive to uninformative ones caused by noises. Moreover, low-rank Rényi’s entropy can be efficiently approximated by random projection and Lanczos iteration techniques, reducing the overall complexity from O(n3) to O(n2s) or even O(ns2), where n is the number of data samples and s n. We conduct large-scale experiments to evaluate the effectiveness of this new information measure, demonstrating superior results compared to matrix-based Rényi’s entropy in terms of both performance and computational efficiency. Copyright © 2022, The Authors. All rights reserved.

关键词： Entropy