|
|
Active Learning for OCR
Kohút, Jan ; Kolář, Martin (referee) ; Hradiš, Michal (advisor)
The aim of this Master's thesis is to design methods of active learning and to experiment with datasets of historical documents. A large and diverse dataset IMPACT of more than one million lines is used for experiments. I am using neural networks to check the readability of lines and correctness of their annotations. Firstly, I compare architectures of convolutional and recurrent neural networks with bidirectional LSTM layer. Next, I study different ways of learning neural networks using methods of active learning. Mainly I use active learning to adapt neural networks to documents that the neural networks do not have in the original training dataset. Active learning is thus used for picking appropriate adaptation data. Convolutional neural networks achieve 98.6\% accuracy, recurrent neural networks achieve 99.5\% accuracy. Active learning decreases error by 26\% compared to random pick of adaptations data.
|
|
|
Application for Sport Group Management
Holcman, Jaroslav ; Tomešek, Jan (referee) ; Kolář, Martin (advisor)
This text describes the process of creating a mobile application for sport club management and its everyday operations. The main goal of the application is to digitize and unite most of common activities in the sport club. Another goal is to create the user interface which will be primarily intuitive and simple to use while providing straightforward ways of usage during most important user use cases. The thesis studies principles for designing user interfaces, analyzes sports team environment and defines the requirements for a software product to support sports club management. The design of the solution follows the analysis of requirements as well as the description of the implementation of the resulting product and its publishing to the real sports club environment. The thesis ends with description of user testing methods and results of the testing which validate accomplishment of defined goals.
|
|
|
Playing Gomoku with Neural Networks
Slávka, Michal ; Kolář, Martin (referee) ; Hradiš, Michal (advisor)
Táto práca sa zaoberá použitím algoritmu AlphaZero pre hru Gomoku. AlphaZero je založený na spätnoväzbnom učení a k trénovaniu nemusia byť využité žiadne existujúce datasety. Trénovanie prebieha iba na hrách algoritmu samého so sebou. AlphaZero používa algoritmus na prehľadávanie stromu, pre zlepšenie stratégie. Na vylepšnej stratégii sa následne trénuje neurónová sieť. Tento prístup bol úspešný v hrách proti existujúcim algoritmom. Generovanie trénovacích dát vysokej kvality si vyžaduje veľa výpočetne náročných iterácií trénovania a generovania dát. Experimenty ukázali, že každou iteráciou sa algoritmus zlepšuje, čo naznačuje, že je ešte miesto na zlepšenie, ale množstvo iterácií nedostačovalo na to, aby bol poriadne natrénovaný.
|
|
|
Detection of Groups and Duplicities of Photos
Václavík, Vojtěch ; Kolář, Martin (referee) ; Zemčík, Pavel (advisor)
The bachelor's thesis aims to create an application that allows import of photographs and to sort duplicities. The application simplifies the import by showing already imported photographs and also hiding the not necessary ones. The application can also be used for finding duplicities or comparing folders, helping the user to clean their photo gallery. Other than that, it is also possible to show photographs with similar timestamps. To identify duplicates the application uses EXIF metadata included in photographs. The application was implemented in C++, graphical interface in Qt framework, work with metadata mediates EasyEXIF library and the database communication uses SQLite library.
|
|
|
Deep Learning Model Uncertainty in Medical Image Analysis
Drevický, Dušan ; Kolář, Martin (referee) ; Kodym, Oldřich (advisor)
Táto práca sa zaoberá určením neistoty v predikciách modelov hlbokého učenia. Aj keď sa týmto modelom darí dosahovať vynikajúce výsledky v mnohých oblastiach počítačového videnia, ich výstupy sú väčšinou deterministické a neposkytujú mnoho informácií o tom, ako si je model istý svojou predpoveďou. To je obzvlášť dôležité pri analýze lekárskych obrazových dát, kde môžu mať omyly vysokú cenu a schopnosť detekovať neisté predikcie by umožnila dohliadajúcemu lekárovi spracovať relevantné prípady manuálne. V tejto práci aplikujem niekoľko rôznych metrík vyvinutých v nedávnom výskume pre určenie neistoty na modely hlbokého učenia natrénované pre lokalizáciu cefalometrických landmarkov. Následne ich vyhodnotím a porovnávam v sade experimentov, ktorých úlohou je určiť, nakoľko jednotlivé metriky poskytujú užitočnú informáciu o tom, ako si je model istý svojou predpoveďou.
|
|
|
Planning System for Healthcare Services
Schauer, Marek ; Beran, Vítězslav (referee) ; Kolář, Martin (advisor)
The information system presented in this work allows planning of nursing care and simplifies its implementation and reporting of care givers. The development of the information system was consulted with a real nursing service located in Bratislava - Karlova Ves. The system offers client records, care giver records, planning of caretaking with the possibility of exceptions and a unique reporting mechanism using geolocation and time stamps. The caretaking system was implemented using web technologies and it is possible to use the system independently of platforms or places. Introduced information system is possible to expand by creating plug-in modules. The goal was to create an application, which is easy to use and has user-friendly experience with the possibility of using in terrain.
|
|
|
Applications of Computers in Fishing Ship
Březina, Zdeněk ; Kolář, Martin (referee) ; Zemčík, Pavel (advisor)
The purpose of this thesis is to examine the problematics of embedded systems for boat fishing and analyzing the existing solutions. The main goal is to design author's own complex solution. The author's solution consists of a mutually communicating Android application and an Arduino-based platform with the necessary sensors.
|
|
|
Playing Games Using Neural Networks
Buchal, Petr ; Kolář, Martin (referee) ; Hradiš, Michal (advisor)
The aim of this bachelor thesis is to teach a neural network solving classic control theory problems and playing the turn-based game 2048 and several Atari games. It is about the process of the reinforcement learning. I used the Deep Q-learning reinforcement learning algorithm which uses a neural networks. In order to improve a learning efficiency, I enriched the algorithm with several improvements. The enhancements include the addition of a target network, DDQN, dueling neural network architecture and priority experience replay memory. The experiments with classic control theory problems found out that the learning efficiency is most increased by adding a target network. In the game environments, the Deep Q-learning has achieved several times better results than a random player. The results and their analysis can be used for an insight to reinforcement learning algorithms using neural networks and to improve the used techniques.
|
|
|
Mobile App for Viewing Internet Pictures
Kočnar, David ; Kolář, Martin (referee) ; Herout, Adam (advisor)
This thesis explains issues concerning the development of mobile applications for Android and describes a complete design and development of mobile application Impres for image viewing. In addition to the mobile application there was also created a server application in PHP providing data from different internet sources. Taken together this service aims to bring to its user an inspiring content consisting of phototography and other visual art. The service will automatically add new content and create user-personalised selection. The work also describes theoretical knowledge and procedures necessary for technical design and development of mobile applications. Conclusion evaluates solution and proposals for possible other features.
|
|
|
Virtual Waiting Room for Medical Practice
Majola, Richard ; Beran, Vítězslav (referee) ; Kolář, Martin (advisor)
The aim of this Bachelor's thesis is to develop virtual waiting room at a doctor as webaplication, made up through development environment Laravel with using PHP languageand MySQL database system. The final aplication allows people to register into the system and log in to waiting queue without having to be physically present in the waiting room.
|
guest ::
RSS feed.