|
|
Recognizing People and Their Activities in Video from Security Cameras
Saloň, Juraj Samuel ; Švec, Tomáš (referee) ; Smrž, Pavel (advisor)
The aim of this thesis is to design and develop a system capable of recognizing the activities of people from surveillance cameras. Special attention is paid to the concept of complex situations or events that are defined by relations between identified objects. The first part surveys state-of-the-art techniques for object recognition, object tracking, and recognition of activities relevant to the realized solution. The second part describes the design and implementation of the devised system. It takes advantage of specific relations among two or more objects that are identified in video recordings, such as "person getting out of the car" or "one or more people met with a person of interest and they left together". Results are evaluated on video data extracted from available datasets and manually annotated. The mean average precision metric (MAP) on the data is reported.
|
|
|
AeroWorks: Visual Identification of Aircraft Flight Regimes
Kardoš, Juraj ; Dittrich, Petr (referee) ; Chudý, Peter (advisor)
This Bachelor thesis deals with the visual identification of an aircraft flight's regimes. It describes the spatial motion of an airplane along with the visualization of flight parameters and also proposes a system for a flight regime visual identification. The system processes the input video on a frame by frame basis in two steps. Initially, the video is being stabilized and the system subsequently proceeds in identification of flight related quantities describing the current flight state. Video stabilization is based on feature points detection and an optical flow calculation. Video frames are transformed in order to achieve sufficient consecutive frames overlap and thus to minimize the parasitic oscillations of the video acquisition system. Identification of values indicated by flight instruments is based on the Hough line transform approach. The thesis also includes a description of an application that analyzes a video from the cockpit of an aircraft and is able to recognize the instrument values displayed on specified flight instruments.
|
|
|
Recognizing and Classification of Traffic Situations
Zbořil, Jiří ; Musil, Petr (referee) ; Smrž, Pavel (advisor)
The aim of this thesis is to identify and classify dangerous situations from surveillance cameras, monitoring traffic. An example of such situations is dangerous standing near by the road and car crash, on which this work focuses. The created system uses object detector, analyzing average images in given interval, K nearest neighbor and K Means algorithm and re-detection of enlarged local area in a frame to select anomaly candidates. Detected objects, that do not belong on the road are eliminated by attaching created road mask. At the very last phase, the interval, together with the classification is determined. Calculated F1 score is 0.645, S4 score 0.535 and precision of classification 80 %.
|
|
|
Shot Boundary Detection
Klicnar, Lukáš ; Herout, Adam (referee) ; Beran, Vítězslav (advisor)
Shot boundary detection is a process of automatically finding the boundaries between shots in a video. This work primarily deals with detection of hard-cuts, but gradual transitions are also considered. At first, fundamental terms of this field are explained, commonly used methods are shortly described. The main part of this work is design and implementation of system for shot boundary detection based on combination of two methods. The first one is comparison of color histograms for adjacent frames. Second approach is based on visual feature tracking. The analysis of behavior of those features is done by estimating their first derivatives. Proposed system was tested on small, manually annotated set of test data, which showed that both features are suitable for this task. Detector proved its ability to find hard-cuts with good precision. It was also able to detect some gradual transitions.
|
|
|
Event detection in camera records
Smolnikov, Mikhail ; Sikora, Pavel (referee) ; Horváth, Tomáš (advisor)
This bachelor thesis focuses on the problem of detection and classification of moving objects in video sequences. The thesis describes the basic algorithms and methods of image data processing, including an introduction to the use of neural networks. The practical part shows the internal logic of a desktop application that allows users to evaluate their own video sequences for the occurrence of movements. The resulting application speeds up the process of video analysis on a selected device many times over.
|
|
|
Recognizing and Classification of Traffic Situations
Zbořil, Jiří ; Musil, Petr (referee) ; Smrž, Pavel (advisor)
The aim of this thesis is to identify and classify dangerous situations from surveillance cameras, monitoring traffic. An example of such situations is dangerous standing near by the road and car crash, on which this work focuses. The created system uses object detector, analyzing average images in given interval, K nearest neighbor and K Means algorithm and re-detection of enlarged local area in a frame to select anomaly candidates. Detected objects, that do not belong on the road are eliminated by attaching created road mask. At the very last phase, the interval, together with the classification is determined. Calculated F1 score is 0.645, S4 score 0.535 and precision of classification 80 %.
|
|
|
Analysis and possibilities of training of selected offensive individual hockey skill in ice hockey in youth categories
Mulač, Jakub ; Perič, Tomáš (advisor) ; Vojta, Zdeněk (referee)
Title: Analysis and possibilities of training of selected offensive individual hockey skill in ice hockey in youth categories Objectives: The first objective was to define the nodal points of the execution of a given movement of a selected offensive individual hockey skill, namely, protecting the puck in ice hockey. The second objective of this thesis was to create skill acqusition program of training of the offensive individual hockey skill of protecting the puck for individual youth category. Methods: The indirect observation method of video was used to obtain the data. A binary yes / no system was used to evaluate the results. Results: The thesis was divided into two parts. The first part was to determine the nodal points of the execution of the selected offensive individual hockey skill protecting the puck in ice hockey using video analysis. A total of nine videos were selected for the video analysis. Three of them contained successful situations of protecting the puck with the body in space, three successful situations of protecting the puck with the body in personal contact and three successful situations of protecting the puck near the boards. First, the nodal points of execution were determined, namely skating, stickhandling and positioning of the puck, and last but not least, body position...
|
|
|
Recognizing People and Their Activities in Video from Security Cameras
Saloň, Juraj Samuel ; Švec, Tomáš (referee) ; Smrž, Pavel (advisor)
The aim of this thesis is to design and develop a system capable of recognizing the activities of people from surveillance cameras. Special attention is paid to the concept of complex situations or events that are defined by relations between identified objects. The first part surveys state-of-the-art techniques for object recognition, object tracking, and recognition of activities relevant to the realized solution. The second part describes the design and implementation of the devised system. It takes advantage of specific relations among two or more objects that are identified in video recordings, such as "person getting out of the car" or "one or more people met with a person of interest and they left together". Results are evaluated on video data extracted from available datasets and manually annotated. The mean average precision metric (MAP) on the data is reported.
|
|
|
Video analysis: an automatic time measurement in the robotic car competition
Ryšlink, Václav ; Vomlelová, Marta (advisor) ; Šikudová, Elena (referee)
Our main goal was to design an algorithm that would automatically evaluate robotic races from the video of the track's finish section. To solve this problem, we used various image processing methods and consequently proposed two different solutions that differ both in the expected input and the inner logic. The first algorithm can evaluate races of cars of an arbitrary appearance since it recognizes cars based on their reference photos. Although this solution proved to be working in all of our experimental recordings, we are aware of a few situations in which this algorithm could be prone to make mistakes. Therefore, we also came up with another algorithm that works more reliably in exchange for demanding cars to have unique color labels. 1
|
|
|
Video analysis for entomology
Šerý, Martin ; Pelikán, Josef (advisor) ; Schier, Jan (referee)
Analysis of behavioral patterns of Drosophila Melanogaster species is a sub- ject of study of many laboratories. The behavior of individuals, for example in different gene lines, is monitored. In order for this analysis to be possible, it is necessary to have a device that is capable of tracking the fly and providing machine-readable data about them. These data are the output of a trace pro- gram FlyCatcher, which records the positions of the observed objects. Thanks to this, we can determine the location, the mutual position of individuals and other parameters. FlyCatcher is based on the use of the AForge library which handles multimedia files. FlyCatcher lets us process videos made in the lab and provide outputs in the .csv format. This format is further processed by external statistical programs.
|
guest ::
