The thesis summarizes the research in application of statistical methods of computational linguistics in music processing and explains theoretical background of these applications. In the second part methods of symbolic melody extraction are explored. A corpus of approxi- mately 400 hours of melodies of different music styles was created. A melody model using the language modeling techniques was trained on this corpus. In the third part of the thesis the model is used for an attempt to develop an alternative method of audio melody extraction which uses the melody model instead of commonly used heuristics and rules. The chosen ap- proach works well only on simple input data and produces worse results than the commonly used methods on the MIREX contest data. On the other hand, the experiments help to understand the conceptual between the pitch frequency development - the physical melody - and the melody perceived on an abstract level in the symbolic notation - the symbolic melody. 1

El presente proyecto de tesis tiene como objetivo desarrollar una aplicación que permita reconocer canciones, por medio del tarareo o canto de una persona. Por este motivo, se ha analizado el estado del arte para conocer cuáles son las mejores soluciones hasta el momento y considerarlos como guía para ofrecer una propuesta. En este proceso, se encontraron los conceptos necesarios para el entendimiento de la problemática y desarrollo de la posible solución. Las partes más importantes del aplicativo son la extracción de descriptores para el poblado de la base de datos y la comparación entre los descriptores y el tarareo para obtener las canciones más similares. / The purpose of this thesis is to develop an application that allows songs to be recognized from the humming or singing of a person. For this reason, the state of the art has been analyzed to know what the best solutions are so far and to consider them as a guide to offer a proposal. In this process, the necessary concepts were found for understanding the problem and developing the possible solution. The most important parts of the application are the extraction of descriptors for the construction of the database and the comparison between the descriptors and the humming to obtain the most similar songs. / Trabajo de investigación