|
|
Camera testing
Frigyik, Tomáš ; Kratochvíla, Lukáš (referee) ; Richter, Miloslav (advisor)
This bachelor's thesis deals with the issue of capturing raster images using cameras and the current state of this issue, the appropriate setting of camera parameters that include external and internal parameters, the design of a measuring set for testing these parameters, while using the designed patterns generated on the monitor. This work also deals with the design of suitable models for setting parameters, the method of their generation and processing. The next part of this work shows a set of images and their processing, evaluation of the properties of the proposed processing procedure and determination of applicability of patterns.
|
|
|
Texture analysis
Opletal, Tomáš ; Kratochvíla, Lukáš (referee) ; Honec, Peter (advisor)
Diploma thesis discusses texture analysis. Goal is to detect anomalies on the material, which could arise during process of making calenders on the nonwoven material. This thesis is using methods of Fourier transform with statistics and Neural networks to detect position of the missing calender.
|
|
|
Object detection
Vítek, Pavel ; Kratochvíla, Lukáš (referee) ; Richter, Miloslav (advisor)
This paper deals with object detection where a human figure is used as a model. The introductory chapter defines terms detection and object. The second chapter describes the basic theory of computer vision , such as image processing and its parts. A substantial part is also focused on the Canny edge detection, upper body detection and pose detection. The third chapter describes the MATLAB integrated development environment in which the image processing functions are implemented. The last chapter is a practical demonstration of object detection. Different upper body detection methods are compared there. In the last subsection, two methods of determining the deviations based on the template are compared. Finally, all methods are compared with respect to the obtained results.
|
|
|
Techniques For Avoiding Model Overfitting On Small Dataset
Kratochvila, Lukas
Building a deep learning model based on small dataset is difficult, even impossible. Toavoiding overfitting, we must constrain model, which we train. Techniques as data augmentation,regularization or data normalization could be crucial. We have created a benchmark with a simpleCNN image classifier in order to find the best techniques. As a result, we compare different types ofdata augmentation and weights regularization and data normalization on a small dataset.
|
|
|
Computing Platforms For Deep Learning Task In Computer Vision
Kratochvila, Lukas
The recent progress in machine learning in computer vision guides to enormous hardware requirements. This paper discovers new innovative hardware capable of dealing with immense demands. The important decision is concentrating on task learning or the final classification. The main concern is on five domains: single-board computers, hardware accelerators, graphics cards, workstations, and cloud computing. These devices have several key features for detection that are discussed. Cloud computing is another presented approach. Furthermore, different delivery models of cloud computing are addressed.
|
|
|
Deep learning model for visual detection and classification general object from industry
Dočkal, Radim ; Honec, Peter (referee) ; Kratochvíla, Lukáš (advisor)
The goal of this bachelor’s thesis is to programme deep learning model for visual detection and classification of general object from industry. The paper is divided into five chapters. First chapter deals with research of the most used architectures of this type. The second chapter deals with choosing the best fitting architecture for usage in industry. In the third chapter is desribed the procedure of creating your own dataset. The fourth chapter then describes the implementation process itself, so that each sub-part of the architecture was sufficiently described and the results are described in the fifht chapter. The summary and recommended procedures for potential implementation in real environment can be found in the conclusion of this paper.
|
|
|
Visual Object Tracking in Realtime
Kratochvíla, Lukáš ; Klečka, Jan (referee) ; Horák, Karel (advisor)
Sledování obecného objektu na zařízení s omezenými prostředky v reálném čase je obtížné. Mnoho algoritmů věnujících se této problematice již existuje. V této práci se s nimi seznámíme. Různé přístupy k této problematice jsou diskutovány včetně hlubokého učení. Představeny jsou reprezentace objektu, datasety i metriky pro vyhodnocování. Mnoho sledovacích algorimů je představeno, osm z nich je implementováno a vyhodnoceno na VOT datasetu.
|
|
|
Li-ion Batery Management System
Kratochvíla, Lukáš ; Macho, Tomáš (referee) ; Kopečný, Lukáš (advisor)
This thesis is about systems for Li-ion batteries. The system consists of measuring and charging. The main part of system is 3 to 5-Series Cell Li-Ion and Li-Phosphate Battery Monitor chip BQ76920. The result of this thesis is printed circuit board (PCB) with BQ76920 and a box contain PCB and battery pack. Box is designed for robot KAMBOT. The system has been tested and measured and evaluated in this thesis.
|
|
|
Li-ion Batery Management System
Kratochvíla, Lukáš ; Macho, Tomáš (referee) ; Kopečný, Lukáš (advisor)
This thesis is about systems for Li-ion batteries. The system consists of measuring and charging. The main part of system is 3 to 5-Series Cell Li-Ion and Li-Phosphate Battery Monitor chip BQ76920. The result of this thesis is printed circuit board (PCB) with BQ76920 and a box contain PCB and battery pack. The system has been tested and measured and evaluated in this thesis.
|
|
| |
guest ::
RSS feed.