|
| |
|
|
4 in 1 Bucket for Loader
Klusáček, David ; Michele, Jiří (referee) ; Škopán, Miroslav (advisor)
The purpose of this master´s thesis is to design multifunctional 4in1 bucket for wheel loader. Develop requested function, size and strenght computation. Documentation of the whole set-up, documentation of sub-parts and documentation of single parts.
|
|
|
Sound Synthesizer Emulator
Kulikov, Maximilian ; Klusáček, David (advisor) ; Kratochvíl, Miroslav (referee)
A tool for creation of emulators of audio synthesizers. The base of the work is an imperative programming language Cynth that describes signals of the resulting sound. Cynth code is translated into C code for further linkage with a program that controls GUI and MIDI input controls and output monitoring and connection with a sound card. The intermediate step of translation into C allows taking advantage of the C compiler optimizations. The Cynth language is restricted in a way that eliminates any dynamic allocations at run-time while allowing complex compile-time programming and working with statically allocated data structures for expressive description of signals. 1
|
|
|
Differentiable Depth Estimation for Bin Picking
Černý, Marek ; Klusáček, David (advisor) ; Šikudová, Elena (referee)
The goal of this thesis was to investigate the neural 3D surface reconstruction from multiple views with the intent to use the resulting depth maps for bin picking. Survey of papers from 2014 to 2018 showed that none of the state of the art methods would be used to control a robot arm in our setup. Therefore we decided to create our low-level neural approach which we called the EmfNet. The network is based on a pyramidal resolution refining approach. At each pyramid's layer, there are three separate networks that take part in the computation. Each of them has a definite goal, which gives us almost complete understanding of what is going on inside the network. The EmfNet model was partially usable, but we nevertheless extended it to EmfNet-v2. First, another measuring layer was added, which freed EmfNet from depending on an unnecessary hyperparameter. Second, we used constraints on geometry for the network not to be confused by occlusions (cases where a certain part of the surface is visible only from a single camera). Both networks were implemented and tested on a corpus that was created as a part of this thesis. A corpus containing rendered as well as real data. The process of correspondence pairing inside the network can be observed using the visualization tool. We designed a way how to use a robotic arm...
|
|
|
Price Explorer
Čížek, Vítězslav ; Kocanda, Jiří (advisor) ; Klusáček, David (referee)
The present work is focused on the problem of processing huge amount of information on the Internet. This task's importance is constantly growing in present days along with the increasing amount of available information. The goal was to develop an application that would solve this problem for one concrete, frequently used area and simplify the user the time-consuming task of data analysis. The program is designed for searching data occurring in on-line shop databases. The technologies used in this application were chosen in a way to satisfy two conditions: to facilitate usage to the end-user and at the same time to enable program's platform independence.
|
|
|
New Methods in Statistical Speech Recognition
Klusáček, David ; Hajič, Jan (advisor) ; Psutka, Josef (referee) ; Černocký, Jan (referee)
Title: New Methods in Statistical Speech Recognition Author: David Klusáček Department: Institute of Formal and Applied Linguistics (ÚFAL), Faculty of Mathematics and Physics in Prague, Malostranské náměstí 25, 118 00 Praha 1. Advisor: Prof. RNDr. Jan Hajič, Dr., Institute of Formal and Applied Linguistics. Abstract: This works aims to identify limits of contemporary speech rec- ognizers and tries to come up with methods that could push back the fron- tiers. After describing the state of the art, the weakest link of the chain has been identified in the acoustic front-end, especially when working in harsh acoustic conditions. NUFIBA front-end, the proposed solution, includes re- verb compensation and speaker/background segmentation as well as contin- uous SNR monitoring which, thru cooperation with acoustic model, hinders from avalanche spreading of recognition errors. Owing to the lack of time, only a phoneme recognizer was finally implemented, although large blocks of originally intended word-based continuous speech recognizer were implemented and tested (such as the MMI-class based language model).
|
|
|
Sound Synthesizer Emulator
Kulikov, Maximilian ; Klusáček, David (advisor) ; Kratochvíl, Miroslav (referee)
A tool for creation of emulators of audio synthesizers. The base of the work is an imperative programming language Cynth that describes signals of the resulting sound. Cynth code is translated into C code for further linkage with a program that controls GUI and MIDI input controls and output monitoring and connection with a sound card. The intermediate step of translation into C allows taking advantage of the C compiler optimizations. The Cynth language is restricted in a way that eliminates any dynamic allocations at run-time while allowing complex compile-time programming and working with statically allocated data structures for expressive description of signals. 1
|
|
| |
|
|
Differentiable Depth Estimation for Bin Picking
Černý, Marek ; Klusáček, David (advisor) ; Šikudová, Elena (referee)
The goal of this thesis was to investigate the neural 3D surface reconstruction from multiple views with the intent to use the resulting depth maps for bin picking. Survey of papers from 2014 to 2018 showed that none of the state of the art methods would be used to control a robot arm in our setup. Therefore we decided to create our low-level neural approach which we called the EmfNet. The network is based on a pyramidal resolution refining approach. At each pyramid's layer, there are three separate networks that take part in the computation. Each of them has a definite goal, which gives us almost complete understanding of what is going on inside the network. The EmfNet model was partially usable, but we nevertheless extended it to EmfNet-v2. First, another measuring layer was added, which freed EmfNet from depending on an unnecessary hyperparameter. Second, we used constraints on geometry for the network not to be confused by occlusions (cases where a certain part of the surface is visible only from a single camera). Both networks were implemented and tested on a corpus that was created as a part of this thesis. A corpus containing rendered as well as real data. The process of correspondence pairing inside the network can be observed using the visualization tool. We designed a way how to use a robotic arm...
|
|
|
Passive emitter tracking
Hrach, Jan ; Klusáček, David (advisor) ; Mareš, Martin (referee)
We have implemented a TDOA multilateration of transmitters on an unmodified rtl-sdr receiver using transmitters with known location as a timing reference. We present a brief theoretical background and describe the measurement process which includes several approaches that correct the timing and frequency errors between the receivers. Additionally, we have implemented an angle of arrival direction finder using coherent rtl-sdr.
|
guest ::
