|
|
Detection and Classification of Road Users in Aerial Imagery Based on Deep Neural Networks
Hlavoň, David ; Hradiš, Michal (referee) ; Rozman, Jaroslav (advisor)
This master's thesis deals with a vehicle detector based on the convolutional neural network and scene captured by drone. Dataset is described at the beginning, because the main aim of this thesis is to create practicly usable detector. Architectures of the forward neural networks which detector was created from are described in the next chapter. Techniques for building a detector based on the naive methods and current the most successful meta architectures follow the neural network architectures. An implementation of the detector is described in the second part of this thesis. The final detector was built on meta architecture Faster R-CNN and PVA neural network on which the detector achieved score over 90 % and 45 full HD frames per seconds.
|
|
| |
|
|
General Processing on Graphics Processing Units for Industrial Systems
Lukačovič, Martin ; Mašek, Jan (referee) ; Krkoš, Radko (advisor)
The thesis deals with the abilities of graphics processors for GPGPU. It contains historical solutions to contemporary design. There are also described graphics processors from the largest manufacturers of this time, their focus and goals in the future. For algorithms implementation using GPU, there are necessary APIs that offer various possibilities of execution. In addition to the CPU and GPU universal heterogeneous computing, there are alternatives such as FPGA and DSP so it is necessary to consider the price and energy cost. Part of the work is devoted to the communication possibilities with the hardware and advanced memory approaches. For demonstrating parallel computing an implementation of matrix multiplication in OpenCL was realized.
|
|
|
Image segmentation on GPU
Bravenec, Tomáš ; Mego, Roman (referee) ; Frýza, Tomáš (advisor)
Bachelor thesis is focused on using graphical processing units for parallel data processing, specifically on image processing. Main focus of this thesis is determining time difference in image processing using graphical processing unit and classic approach on processor. Another focus is accessing webcam and processing of captured frames.
|
|
|
Paralelized image processing library
Fuksa, Tomáš ; Macho, Tomáš (referee) ; Petyovský, Petr (advisor)
This work deals with parallel computing on modern processors - multi-core CPU and GPU. The goal is to learn about computing on this devices suitable for parallelization, define their advantages and disadvantages, test their properties in examples and select appropriate tools to implement a library for parallel image processing. This library is going to be used for the vanishing point estimation in the path finding mobile robot.
|
|
|
Object Detection on GPU
Macenauer, Pavel ; Polok, Lukáš (referee) ; Juránek, Roman (advisor)
This thesis addresses the topic of object detection on graphics processing units. As a part of it, a system for object detection using NVIDIA CUDA was designed and implemented, allowing for realtime video object detection and bulk processing. Its contribution is mainly to study the options of NVIDIA CUDA technology and current graphics processing units for object detection acceleration. Also parallel algorithms for object detection are discussed and suggested.
|
|
|
Deep neural network for supercomputer environments
Bronda, Samuel ; Kolařík, Martin (referee) ; Burget, Radim (advisor)
The main benefit of the work is the optimization of the hardware configuration for the calculation of neural networks. The theoretical part describes neural networks, deep learning frameworks and hardware options. The next part of the thesis deals with implementation of performance tests, which include application of Inception V3 and ResNet models. Network models are applied to various graphics cards and computing hardware. The output of the thesis is the implemented model of the network Inception V3, which examines the graphics cards and their performance, time-consuming calculations and their efficiency. The ResNet model is applied to a section that examines other impacts on neural network computing such as used disk, operating memory, and so on. Each practical part contains a discussion where the knowledge of the given part is explained. In the case of consumption measurement, a mismatch between the declaration by the manufacturer and the measured values was identified.
|
|
|
Programming CUDA with C#
Zajíc, Jiří ; Kačic, Matej (referee) ; Jurnečka, Peter (advisor)
This master's thesis is focused on GPU accelerated calculations on NVidia graphics card. CUDA technology is used and converted to implementation on a .NET platform. The problem is solved as a compiler from C# programing language to NVidia CUDA language with expression atrributes of C# language that preserves the same semantics of actions. Application is implemented in C# programing language and uses NRefactory, the open-source library.
|
|
|
Acceleration of Ultrasound Neurostimulation Using Mixed-Precision Arithmetic
Duchoň, Radek ; Olšák, Ondřej (referee) ; Jaroš, Jiří (advisor)
K-Wave is an open source tool for acoustic and ultrasound simulation. Current available implementations are written in C++ and Matlab. The aim of this thesis is to accelerate the existing implementation of ultrasound simulation by means of lower precision calculations on Nvidia graphics cards using the CUDA software platform. Another benefit of this work should be a reduced memory requirement, which will enable larger simulations to be performed. However, due to the use, for example, for neurostimulation of the brain, the reduced accuracy must not disturb the results as a whole too much. It is therefore important to identify suitable quantities that can be stored in lower precision. In this work, possible approaches and their effectiveness in utilizing lower precision will be analyzed. Furthermore, a solution proposal will be made, which will include identifying potential variables for reduction. This will be followed by specifying the achieved implementation and its testing. The conclusion will focus on evaluating the solution based on the results obtained from the testing.
|
|
|
Image Segmentation On CPU/GPU
Bravenec, Tomáš
This article is focused on using graphical processing units for parallel data processing, specifically on image processing. Main focus of this thesis is determining time difference in image processing using graphical processing unit and classic approach on processor. Another focus is accessing webcam and processing of captured frames.
|
guest ::
