|
|
Estimation and analysis of speech signal periodicity
Malucha, Jan ; Poměnková, Jitka (referee) ; Sigmund, Milan (advisor)
Theoretical part starts with a short introduction to speech signals and short time analysis method. Concept of periodicity is clarified. This is followed by analysis of the speech signal parameters - voiced/unvoiced classification, intonation and short time period similarity. Next the overview of methods used to analyze the mentioned parameters together with concepts of their algorithms is provided. These methods include ZCR, STE, HNR, ACF, NCC, AMDF and DTW. Algorithms using the methods with supporting preprocessing algorithms were programmed in MATLAB for practical part. All were tested on real speech signals and the results are discussed at the end of the thesis.
|
|
|
Searching for Similar 3D Models
Karásek, Miroslav ; Najman, Pavel (referee) ; Španěl, Michal (advisor)
The work deals with the design and implementation of similarity based search in database of 3D models. The database is realised using of quick Python data structures. For Spherical harmonics and 3D Zernike deskriptors extraction are used third-party libraries. REST application interface is designed to access the search engine. A web based application was designed as an example of the REST application interface usage. Furthermore, the work deals with the search quality, presents results of the implemented database and suggests possible improvements.
|
|
| |
|
|
Sound records comparison using timbre features
Miklánek, Štěpán ; Schimmel, Jiří (referee) ; Kiska, Tomáš (advisor)
This thesis deals with research of musical features, which are describing music recordings relating to timbre. First chapter deals with historical development and modern approach in a discipline called Music Information Retrieval (MIR), further there is a description of music processing from the perspective of music theory and digital signal processing. Then followed by a description of signal pre-processing. This part is very important when retrieving features from music recordings. In chapter concerned about retrieving features there are summarized all common features used when retrieving information from musical recordings with main concern to timbral features. A database of music recordings and a feature retrieving system is introduced. The last chapter deals with individual analysis of timbral features.
|
|
|
Music recommendation based on music information retrieval
Semela, René ; Schimmel, Jiří (referee) ; Kiska, Tomáš (advisor)
This thesis deals with the design, implementation and testing of the content-based music recommender system based on music information retrieval. In the introduction the attention is paid to issues of music information retrieval and to areas of their utilization, it also focuses on tools of their retrieving. Aferwards the most used types of recommender systems are described, including their typical problems. Options of the hybridization of these systems as well as examples of the popular music recommender systems are mentioned. There also is an outline of their functioning. The following section is focused on the parameterization of musical pieces and is devoted to the description of particular most used parameters. The next section is devoted to the content-based music recommender system design itself, including the defining of particular parameters that are used to differentiate musical recordings using the algorithm mRMR and other procedures. The recommender system design as such is oriented to the classification method k-nearest neighbors. The attention is also paid to the model of user taste recorded by Rocchio algorithm. In the next section the system is implemented according to the design. There is also described its functionality including the background processes. The final part of this work is focused on system testing and evaluation.
|
|
|
Searching for Similar 3D Models
Sálus, Ondřej ; Beran, Vítězslav (referee) ; Španěl, Michal (advisor)
Goal of this paper is development of aplication for comparing of 3D models. It deals with extraction od descriptors suitable for comparing of 3D models and enables to find models similar to users query. For extraction of descriptors application uses Reeb graph method. This paper also describes usage of libraries that implements Reeb graph method and some experiments with speed and quality over pre prepared models database.
|
|
|
Image similarity measuring using deep learning
Štarha, Dominik ; Šeda, Pavel (referee) ; Rajnoha, Martin (advisor)
This master´s thesis deals with the reseach of technologies using deep learning method, being able to use when processing image data. Specific focus of the work is to evaluate the suitability and effectiveness of deep learning when comparing two image input data. The first – theoretical – part consists of the introduction to neural networks and deep learning. Also, it contains a description of available methods, their benefits and principles, used for processing image data. The second - practical - part of the thesis contains a proposal a appropriate model of Siamese networks to solve the problem of comparing two input image data and evaluating their similarity. The output of this work is an evaluation of several possible model configurations and highlighting the best-performing model parameters.
|
|
| |
|
|
A tool for simultaneous playback of multiple composition interpretations
Švejcar, Michael ; Ištvánek, Matěj (referee) ; Miklánek, Štěpán (advisor)
The purpose of this Bachelor’s thesis was to create a piece of software which enables the user to simultaneously play back multiple interpretations of a musical piece and switch between them instantaneously. This was achieved using the App Designer in the MATLAB environment, which is intended for developing applications with graphical user interface. The key to the development of the application was especially the use of available toolboxes and algorithms for computing chromagrams and multiscale dynamic time warping. The final IntSwitcher player enables the user to load two recordings of interpretations of one song. Chromagrams which characterize the individual recordings in terms of tonal development over time are first calculated from the input files. After that, the multiscale dynamic time warping method is applied on the chromagrams, which outputs the warping path. The warping path in this case is a matrix, in which musically corresponding samples of loaded audio files are assigned together with the resolution of 50 ms. From this, the corresponding time position of currently inactive track is computed along with its slider position. If the user switches the currently played recording, the second track starts playing in the same part of composition, even if that part is at a different time in each of the individual recordings. The final software is an appropriate tool for studying differences between various interpretations of the same musical piece.
|
|
|
MediaDiff - Diff for Static Images
Brothánek, Jan ; Mlích, Jozef (referee) ; Chmelař, Petr (advisor)
This bachelor's thesis focuses on comparison of digital documents, especially static images. This paper presents algorithms, which are usable for general data comparison and also specific algorithms major in image data comparison. We introduce possibilities of document comparison which are now available for users. On the basis of this current situation, a prototype of new modular application MediaDiff for various documents comparison is designed and implemented. We describe in detail each stage of the application development, problems which had arisen and their solutions. The possibilities of future development of the application are discussed.
|
guest ::
