|
Acceleration of Object Detection Using Classifiers
Juránek, Roman ; Kälviäinen, Heikki (referee) ; Sojka, Eduard (referee) ; Zemčík, Pavel (advisor)
Detekce objektů v počítačovém vidění je složítá úloha. Velmi populární a rozšířená metoda pro detekci je využití statistických klasifikátorů a skenovacích oken. Pro učení kalsifikátorů se často používá algoritmus AdaBoost (nebo jeho modifikace), protože dosahuje vysoké úspěšnosti detekce, nízkého počtu chybných detekcí a je vhodný pro detekci v reálném čase. Implementaci detekce objektů je možné provést různými způsoby a lze využít vlastnosti konkrétní architektury, pro urychlení detekce. Pro akceleraci je možné využít grafické procesory, vícejádrové architektury, SIMD instrukce, nebo programovatelný hardware. Tato práce představuje metodu optimalizace, která vylepšuje výkon detekce objektů s ohledem na cenovou funkci zadanou uživatelem. Metoda rozděluje předem natrénovaný klasifikátor do několika různých implementací, tak aby celková cena klasifikace byla minimalizována. Metoda je verifikována na základním experimentu, kdy je klasifikátor rozdělen do předzpracovací jednotku v FPGA a do jednotky ve standardním PC.
|
|
Sharing Local Information for Faster Scanning-Window Object Detection
Hradiš, Michal ; Kälviäinen, Heikki (referee) ; Matas, Jiří (referee) ; Zemčík, Pavel (advisor)
Cílem této dizertační prace je vylepšit existující detektory objektů pomocí sdílení informace a výpočtů mezi blízkými pozicemi v obraze. Navrhuje dvě metody, které jsou založené na Waldově sekvenčním testu poměrem pravděpodobností a algoritmu WaldBoost. První z nich, Early non-Maxima Suppression , přesunuje rozhodování o potlačení nemaximálních pozic ze závěrečné fáze do fáze vyhodnocování detektoru, čímž zamezuje zbytečným výpočtům detektoru v nemaximálních pozicích. Metoda neighborhood suppression doplňuje existující detektory o schopnost zavrhnout okolní pozice v obraze. Navržené metody je možné aplikovat na širokou škálu detektorů. Vyhodnocení obou metod dokazují jejich výrazně vyšší efektivitu v porovnání s detektory, které vyhodnocují jednotlivé pozice obrazu zvlášť. Dizertace navíc prezentuje výsledky rozsáhlých experimentů, jejichž cílem bylo vyhodnotit vlastnosti běžných obrazových příznaků v několika detekčních úlohách a situacích.
|
|
Automatic Machine Learning Methods for Multimedia Data Analysis
Mašek, Jan ; Chromý, Erik (referee) ; Vozňák, Miroslav (referee) ; Burget, Radim (advisor)
The quality and efficient processing of increasing amount of multimedia data is nowadays becoming increasingly needed to obtain some knowledge of this data. The thesis deals with a research, implementation, optimization and the experimental verification of automatic machine learning methods for multimedia data analysis. Created approach achieves higher accuracy in comparison with common methods, when applied on selected examples. Selected results were published in journals with impact factor [1, 2]. For these reasons special parallel computing methods were created in this work. These methods use massively parallel hardware to save electric energy and computing time and for achieving better result while solving problems. Computations which usually take days can be computed in minutes using new optimized methods. The functionality of created methods was verified on selected problems: artery detection from ultrasound images with further classifying of artery disease, the buildings detection from aerial images for obtaining geographical coordinates, the detection of materials contained in meteorite from CT images, the processing of huge databases of structured data, the classification of metallurgical materials with using laser induced breakdown spectroscopy and the automatic classification of emotions from texts.
|
|
Speed Measurement Using Radar
Andrla, Jiří ; Široký, Adam (referee) ; Maršík, Lukáš (advisor)
The principal goal of this bachelor thesis is to design a program in the Matlab programming language that measures the speed of vehicles scanned by a radar. In this thesis, the program is also implemented in the C++ programming language. The designed software detects moving vehicles in sets of data obtained from a CW radar. For each of these vehicles, their speed is computed with respect to the angle of their motion to the orientation of the measuring device. In the end, overall statistics are generated from the data gained. The thesis solves the problems of correction of the speed measurement based on the measuring angle. This technique should provide better results than other classical methods without the correction. The outcome of this thesis can be interesting for every application or system that uses speed measurement with a radar under different angles.
|
|
Manipulace s objekty pomocí robotu Mitsubishi RV2AJ založená na analýze obrazu
Koutský, Filip
This Thesis deals with proposal, implementation and testing of a program for detection of objects to be manipulated and a program for control of robotic arm Melfa RV-2AJ manipulating with these objects. The Theoretical part describes selected algorithms for object detection using a Basler camera and introduces options for controlling the robotic arm Melfa RV-2AJ. The Practical part focuses on development and implementation of a program in the LabVIEW Development Environment for object detection using LabVIEW module NI Vision Acquisition Software and NI Vision Builder for Automated Inspection. Furthermore it describes the development of a program for control of robotic arm Melfa RV-2AJ with use of the LabVIEW module for Mitsubishi robot control called Imaging Lab. The program has been tested by a set of testing tasks and the test results subsequently used for suggestions for functionality improvements.
|
|
Robocar - The Autonomous Driving
Hnát, Miroslav ; Richter, Miloslav (referee) ; Petyovský, Petr (advisor)
In my Bachelor thesis I deal with automated car driving requirements. First I make the real plan and select the correct software tool for simulation and realization this issues. Second I solve model for autonomous car driving simulation, breaking and rotation wheels. There is open area with several obstacles in simulation conditions. I define possibly errors and I recommend its solution. Next very important part of my thesis are identify the parameters of real vehicle and verification the correctness of model by simulation test. I review variances from model. The object of this has to check reactions of the real system. Whole solution will be implemented into individual components. The individual components will be then compared with simulation results. In conclusion I will appreciate achieved results and recommend my solutions.
|
|
Image classification using deep learning
Hřebíček, Zdeněk ; Přinosil, Jiří (referee) ; Mašek, Jan (advisor)
This thesis deals with image object detection and its classification into classes. Classification is provided by models of framework for deep learning BVLC/Caffe. Object detection is provided by AlpacaDB/selectivesearch and belltailjp/selective_search_py algorithms. One of results of this thesis is modification and usage of deep convolutional neural network AlexNet in BVLC/Caffe framework. This model was trained with precision 51,75% for classification into 1 000 classes. Then it was modified and trained for classification into 20 classes with precision 75.50%. Contribution of this thesis is implementation of graphical interface for object detction and their classification into classes, which is implemented as aplication based on web server in Python language. Aplication integrates object detection algorithms mentioned abowe with classification with help of BVLC/Caffe. Resulting aplication can be used for both object detection (and classification) and for fast verification of any classification model of BVLC/Caffe. This aplication was published on server GitHub under license Apache 2.0 so it can be further implemented and used.
|
|
RoboAuto - Detection of Moving Objects
Štibinger, Petr ; Žák, Pavel (referee) ; Juránek, Roman (advisor)
In this work is possible to find out something about detection of moving objects from moving camera. Method that was used is MTI (moving target indication), which is used for detection from air and can be applied on ground vehicles. Introduced will be work with SURF (for detection of interesting points in image) and histograms (for detection moving objects).
|
|
Object Detection in Images
Ptáček, Tomáš ; Šiler, Ondřej (referee) ; Švub, Miroslav (advisor)
This work deals with the problem of object detection in images and describes theoretical backgrounds of detection based on boosting, AdaBoost algorithm and Haar-like features as weak classifiers. Further this work engages in design and implementation of a training and detection application based on OpenCV and wxWidgets libraries. To the end it shows a training and face detection test performed in the implemented application.
|
|
License Plate Recognition
Tilňak, Tomáš ; Juránek, Roman (referee) ; Herout, Adam (advisor)
This thesis talks about a license plate recognition problematics and my implementation of license plate recognition program. At first I introduce a format of license plates in Czech republic. Next chapter is about existing solutions for each phase of license plate recognition according to the selected scientific articles. The main part of this thesis is about design and implementation of license plate recognition program. I also introduce libraries I used in implementation. Necessary part in software development is testing, which has also its own chapter. In the final part there is a review of results and proposals for future changes.
|