scispace - formally typeset
Search or ask a question
Institution

Brno University of Technology

EducationBrno, Czechia
About: Brno University of Technology is a education organization based out in Brno, Czechia. It is known for research contribution in the topics: Fracture mechanics & Filter (video). The organization has 6339 authors who have published 15226 publications receiving 194088 citations. The organization is also known as: Vysoké učení technické v Brně & BUT.


Papers
More filters
Proceedings ArticleDOI
01 Dec 2011
TL;DR: This work describes how to effectively train neural network based language models on large data sets and introduces hash-based implementation of a maximum entropy model, that can be trained as a part of the neural network model.
Abstract: We describe how to effectively train neural network based language models on large data sets. Fast convergence during training and better overall performance is observed when the training data are sorted by their relevance. We introduce hash-based implementation of a maximum entropy model, that can be trained as a part of the neural network model. This leads to significant reduction of computational complexity. We achieved around 10% relative reduction of word error rate on English Broadcast News speech recognition task, against large 4-gram model trained on 400M tokens.

539 citations

Journal ArticleDOI
TL;DR: Supercapacitors represent the alternative to common electrochemical batteries, mainly to widely spread lithium-ion batteries as discussed by the authors, and their properties are between batteries and capacitors, where they are able to quickly accommodate large amounts of energy (smaller than in the case of batteries).
Abstract: Energy accumulation and storage is one of the most important topics in our times. This paper presents the topic of supercapacitors (SC) as energy storage devices. Supercapacitors represent the alternative to common electrochemical batteries, mainly to widely spread lithium-ion batteries. By physical mechanism and operation principle, supercapacitors are closer to batteries than to capacitors. Their properties are somewhere between batteries and capacitors. They are able to quickly accommodate large amounts of energy (smaller than in the case of batteries – lower energy density from weight and volume point of view) and their charging response is slower than in the case of ceramic capacitors. The most common type of supercapacitors is electrical double layer capacitor (EDLC). Other types of supercapacitors are lithium-ion hybrid supercapacitors and pseudo-supercapacitors. The EDLC type is using a dielectric layer on the electrode − electrolyte interphase to storage of the energy. It uses an electrostatic mechanism of energy storage. The other two types of supercapacitors operate with electrochemical redox reactions and the energy is stored in chemical bonds of chemical materials. This paper provides a brief introduction to the supercapacitor field of knowledge.

427 citations

Proceedings ArticleDOI
15 Apr 2007
TL;DR: This work is exploring the possibility of obtaining the features directly from neural net without the necessity of converting output probabilities to features suitable for subsequent GMM-HMM system.
Abstract: In recent years, probabilistic features became an integral part of state-of-the-are LVCSR systems. In this work, we are exploring the possibility of obtaining the features directly from neural net without the necessity of converting output probabilities to features suitable for subsequent GMM-HMM system. We experimented with 5-layer MLP with bottle-neck in the middle layer. After training such a neural net, we used outputs of the bottle-neck as features for GMM-HMM recognition system. The benefits are twofold: first, improvement was gained when these features are used instead of the probabilistic features, second, the size of the system was reduced, as only part of the neural net is used. The experiments were performed on meetings recognition task defined in MST RT'05 evaluation.

389 citations

Journal ArticleDOI
TL;DR: The concept of matched filtering is improved, and the proposed blood vessel segmentation approach is at least comparable with recent state-of-the-art methods, and outperforms most of them with an accuracy of 95% evaluated on the new database.
Abstract: Automatic assessment of retinal vessels plays an important role in the diagnosis of various eye, as well as systemic diseases. A public screening is highly desirable for prompt and effective treatment, since such diseases need to be diagnosed at an early stage. Automated and accurate segmentation of the retinal blood vessel tree is one of the challenging tasks in the computer-aided analysis of fundus images today. We improve the concept of matched filtering, and propose a novel and accurate method for segmenting retinal vessels. Our goal is to be able to segment blood vessels with varying vessel diameters in high-resolution colour fundus images. All recent authors compare their vessel segmentation results to each other using only low-resolution retinal image databases. Consequently, we provide a new publicly available high-resolution fundus image database of healthy and pathological retinas. Our performance evaluation shows that the proposed blood vessel segmentation approach is at least comparable with recent state-of-the-art methods. It outperforms most of them with an accuracy of 95% evaluated on the new database.

371 citations

01 Jun 2010
TL;DR: In this article, the authors give an overview of the challenges a system designer has to consider while implementing an FSO system, including typical gains and losses along the path from the transmitter through the medium to the receiver.
Abstract: Over the last two decades free-space optical communication (FSO) has become more and more interest- ing as an adjunct or alternative to radio frequency commu- nication. This article gives an overview of the challenges a system designer has to consider while implementing an FSO system. Typical gains and losses along the path from the transmitter through the medium to the receiver are in- troduced in this article. Detailed discussions of these topics can be found in this special issue of the Radioengineering Journal.

333 citations


Authors

Showing all 6383 results

NameH-indexPapersCitations
Georg Kresse111430244729
Patrik Schmuki10976352669
Michael Schmid8871530874
Robert M. Malina8869138277
Jiří Jaromír Klemeš6456514892
Alessandro Piccolo6228414332
René Kizek6167216554
George Danezis5920911516
Stevo Stević583749832
Edvin Lundgren5728610158
Franz Halberg5575015400
Vojtech Adam5561114442
Lukas Burget5325221375
Jan Cermak532389563
Hynek Hermansky5131714372
Network Information
Related Institutions (5)
Vienna University of Technology
49.3K papers, 1.3M citations

87% related

Polytechnic University of Catalonia
45.3K papers, 949.3K citations

86% related

Fraunhofer Society
40.1K papers, 820.8K citations

86% related

Polytechnic University of Milan
58.4K papers, 1.2M citations

86% related

Aalto University
32.6K papers, 829.6K citations

85% related

Performance
Metrics
No. of papers from the Institution in previous years
YearPapers
202328
2022106
20211,053
20201,010
20191,214
20181,131