|
|
Identification of Persons in the Video from Quadcopter
Mojžiš, Tomáš ; Drahanský, Martin (referee) ; Goldmann, Tomáš (advisor)
The aim of this thesis is to make an application capable of recognizing people's faces based on a user-created database in drone footage. The database is made of pictures of people that should be identified in the footage. The output of this application is a video where the demanded people are labeled with their names. Some face detection and recognition state of the art solutions based on neural networks are compared in this work. The final solution consists of the MTCNN detector and a face embedding extractor based on ArcFace. The created multiplatform application allows to recognize people in drone footage even with face width of less than 20 pixels. The final solution was tested on a private dataset comprised of drone footage.
|
|
|
Face Features Extraction Methods
Adamček, Ľubomír ; Orság, Filip (referee) ; Goldmann, Tomáš (advisor)
The face has been one of the most attractive available human biometries for a long time due to it's easy and convenient way to obtain it. Possibilities of it's utilization are broad - from security, through monitoring up to enternainment industry. This work presents the domain of face biometry and analyses 3 extraction methods of facial features - PCA, LBP and HOG. A part of this work is also an efficient implementation of these algorithms including GUI designed to experiment with this implementation and to evaluate its performance on a set of images capturing people in various conditions.
|
|
|
Virtual Robot Control Using EEG
Drla, Michal ; Goldmann, Tomáš (referee) ; Tinka, Jan (advisor)
This bachelor thesis aimed to create an application where is user able to control the virtual robot with an EEG signal. The thesis contains a brief introduction that explains how BCI systems which are using EEG work. This introduction not only explains the basics of EEG analysis but also explains brain biology and shows different signals which are extractable from the brain. This thesis also explains the theory of neural networks which are used to implement the analysis. In implementation are shown scripts that were used to collect data and there is also shown the design of the neural network. Results of testing are good, the neural network was making correct decisions and the user was able to control the virtual robot.
|
|
|
Reconstruction of Facial Images Using Neural Networks
Zubalík, Petr ; Drahanský, Martin (referee) ; Goldmann, Tomáš (advisor)
The main purpose of this bachelor's thesis is to propose and implement a model, using neural networks, that will be able to reconstruct low-resolution facial images with blurry parts of the face. The task of super-resolution of facial images is solved by two models based on convolutional neural networks. The first model is built upon the architecture of ResNet whereas the other model makes use of the principles of generative adversarial networks. The proposed models are implemented in the Python programming language with the use of application programming interface of the TensorFlow framework. Moreover, as a part of this work, an application with a simple grafical user interface was created. This application makes it easy to use the implemented models. Several experiments are analyzed in the last chapter of this thesis to evaluate the performance of the models.
|
|
|
Detection of Anomalies in Pedestrian Walking
Pokorný, Ondřej ; Orság, Filip (referee) ; Goldmann, Tomáš (advisor)
The goal of this work was to create a system that would be able to detect anomalies in pedestrian walking. As the core of my application, I have used OpenPose, which is an application for detecting human skeletons. Then I used a bidirectional LSTM neural network to detect anomalies in video sequences. This architecture was chosen during the experiment because it outperformed other solutions. I trained my model to detect three types of anomalies. The output of my application is a video with marked sequences of anomalies. The whole system is implemented in Python.
|
|
|
Estimate the Height of a Persons from the Video Data
Jelen, Vilém ; Orság, Filip (referee) ; Goldmann, Tomáš (advisor)
The aim of this bachelor thesis is to create an application that can detect pedestrians and make an estimate of their height from the video data. In the main part of the thesis I introduced security cameras and image processing methods suitable for detecting and classifying objects. Based on the comparison of these methods, I realized the pedestrian detection using Histograms of Oriented Gradients. As the input parameters of the proposed height estimation algorithm, I used the position and properties of the camera and reference objects or points.
|
|
|
Car Localization Using Satellite Systems
Uhliar, Martin ; Drahanský, Martin (referee) ; Goldmann, Tomáš (advisor)
This bachelor thesis deals with the design, hardware and software implementation of a location device which can get a location data and speed of a car on the Earth's surface. Through the GPRS mobile network, the obtained values are transferred to a remote server and then stored in a database. This data is presented to the user via a web application that displays the current location of the car on the map. The device is controlled by ARM microcontroller.
|
|
|
Face Detection in Poor Quality Videos
Koval, Michal ; Orság, Filip (referee) ; Goldmann, Tomáš (advisor)
This bachelor thesis deals with face detection in low quality videos, while mainly focusing on occluded faces. It describes elementary priciples of machine learning algorithms and their methods, which are often used in the field of computer vision. Out of them are more closely described convolutional neural networks and their state of the art models focused on face detection. Out of those, convolutional neural networks and state of the art models for face detection are more closely described. For the practical part face detection models inspired by state of the art model RetinaFace were implemented and trained. The best performing model achieves 85.5% average precision on WIDER Face HARD testing dataset and 90.9% on dataset focused on occluded faces. Part of this thesis is also a program with graphical user interfaces which provides tools to use developed models on videos and pictures.
|
|
|
Determination of Gun Type and Position in Image Scene
Kolcún, Róbert ; Goldmann, Tomáš (referee) ; Drahanský, Martin (advisor)
The main goal of this work is to design an approach for classifying guns into two categories, with short and long weapons, and determining a position of guns in the image scene. This problem was solved using the classical approach as K-Nearest-Neighbour and SVM classification and using convolutional neural networks. In this work was made a comparison of these approaches with a resulting accuracy up to 90%. With the results of this work, it is possible to choose a right approach to this problem.
|
|
|
Monitoring Pedestrian by Drone
Dušek, Vladimír ; Goldmann, Tomáš (referee) ; Drahanský, Martin (advisor)
This thesis is focused on monitoring people in a video footage captured by drone. People are detected by trained model of detector RetinaNet. A feature vector is extracted for each detected person using color histograms. Identification of people is realized by comparing their feature vectors with respect to their distance in the frame. In the end the trajectories of all people are visualized in a panorama image. Accuracy of the trained RetinaNet detector on difficult validation data is 58.6 %. Error rate is partially reduced by the way of algorithm design for trajectory visualisation. It's not necessary to successfully detect person on every frame for correct visualization of its trajectories. At the same time, static objects which are detected as person but are not moving are not consider as people and are not visualized at all. There is a lot of algorithms dealing with people detection however only a few approaches are focused on detection people from an aerial footage.
|
guest ::
RSS feed.