|
|
Clean photo out of corrupted videosequence
Berky, Martin ; Záviška, Pavel (referee) ; Rajmic, Pavel (advisor)
This diploma thesis deals with separation of moving objects from static unchanging background in video sequence. Thesis contains description of common method of separation and approach based sparse signal representation. In the practical part of thesis, there were created video sequences, which are used to verify designed algorithm implemented in Matlab interface, disegned to obtain separated background from damaged video frames.
|
|
|
Two video-processing problems by means of nontraditional methods
Kánský, Antonín ; Mangová, Marie (referee) ; Rajmic, Pavel (advisor)
The aim of this work is to solve two problems from the field of video editing by means of sparse representation of signals. The problematics of the traditional realisation of two effects, which are background separation of and separation background from moving foreground, is clarified here, as well as the problematics of sparse signals. The solutions was achieved through the method of Principal component analysis (PCP). The resulting algorithm is implemented and tested by simulated and real data.
|
|
|
Separation of dynamic and static structures in a series of images
Gebrtová, Karolína ; Kosová, Petra (referee) ; Rajmic, Pavel (advisor)
This bachelor thesis is focused on the separation of static and dynamic structures of video. While the major part of the video background remains static, a few moving particles can be observed. Such video can be represented by a low-rank and a sparse component. Using a low-rank structure, the background can be separated by applying either the median filter or the dynamic mode decomposition. Furthermore, the thesis includes the robust principal component analysis which can be perceived as an optimization problem. In addition, the thesis is enhanced by result visualization methods.
|
|
|
Clean photo out of corrupted videosequence
Berky, Martin ; Záviška, Pavel (referee) ; Rajmic, Pavel (advisor)
This diploma thesis deals with separation of moving objects from static unchanging background in video sequence. In this thesis are described common method of separation and access using sparse signal representation. In the practical part of thesis was created video sequences, on which is verified the designed algorithm, implemented in Matlab, for obtaining background from damaged video frames and comparing this methods.
|
|
|
Decomposition of video-sequences into components with different dynamics
Gebrtová, Karolína ; Druckmüller, Miloslav (referee) ; Rajmic, Pavel (advisor)
The thesis is focused on the decomposition of video sequences, primarily on background and dynamic component separation, where the background remains the same and only small parts are in motion. Such a video is represented by a low-rank and sparse component. Thanks to the low-rank structure, the background can be separated using a median filter and dynamic mode decomposition. Furthermore, the robust principal component analysis, formulated as an optimization problem and its non-convex variant, will be employed. Also mentioned will be the multiresolution DMD, which is capable of decomposing the dynamic component based on its velocity.
|
|
|
Separation of dynamic and static structures in a series of images
Gebrtová, Karolína ; Kosová, Petra (referee) ; Rajmic, Pavel (advisor)
This bachelor thesis is focused on the separation of static and dynamic structures of video. While the major part of the video background remains static, a few moving particles can be observed. Such video can be represented by a low-rank and a sparse component. Using a low-rank structure, the background can be separated by applying either the median filter or the dynamic mode decomposition. Furthermore, the thesis includes the robust principal component analysis which can be perceived as an optimization problem. In addition, the thesis is enhanced by result visualization methods.
|
|
|
Clean photo out of corrupted videosequence
Berky, Martin ; Záviška, Pavel (referee) ; Rajmic, Pavel (advisor)
This diploma thesis deals with separation of moving objects from static unchanging background in video sequence. Thesis contains description of common method of separation and approach based sparse signal representation. In the practical part of thesis, there were created video sequences, which are used to verify designed algorithm implemented in Matlab interface, disegned to obtain separated background from damaged video frames.
|
|
|
Clean photo out of corrupted videosequence
Berky, Martin ; Záviška, Pavel (referee) ; Rajmic, Pavel (advisor)
This diploma thesis deals with separation of moving objects from static unchanging background in video sequence. In this thesis are described common method of separation and access using sparse signal representation. In the practical part of thesis was created video sequences, on which is verified the designed algorithm, implemented in Matlab, for obtaining background from damaged video frames and comparing this methods.
|
|
|
Two video-processing problems by means of nontraditional methods
Kánský, Antonín ; Mangová, Marie (referee) ; Rajmic, Pavel (advisor)
The aim of this work is to solve two problems from the field of video editing by means of sparse representation of signals. The problematics of the traditional realisation of two effects, which are background separation of and separation background from moving foreground, is clarified here, as well as the problematics of sparse signals. The solutions was achieved through the method of Principal component analysis (PCP). The resulting algorithm is implemented and tested by simulated and real data.
|
guest ::
