Dissertation > Excellent graduate degree dissertation topics show

HMM model humming

Author: ShaXiaoYan
Tutor: GengGuoHua
School: Northwestern University
Course: Computer Software and Theory
Keywords: The Music feature extraction and expression Hidden Markov Models Query by Humming
CLC: TP391.3
Type: Master's thesis
Year: 2008
Downloads: 180
Quote: 1
Read: Download Dissertation

Abstract


The computer retrieve audio clips , text annotation methods can be used based on the title or file name , but it is difficult to find the file name and a textual description of the incompleteness and subjectivity , to meet the specific requirements of the audio clip . Humming music retrieval as a most natural way music retrieval has domestic and foreign experts widespread concern . HMM statistical model as a solid theoretical foundation , has been introduced humming music retrieval . But most of the research is to note based model , for the processing unit to carry out training and recognition of musical notes , notes segmentation to obtain the sound of each note of the melody long , this will cause the humming unnatural way . To solve the above problem , this paper humming retrieval theory and the HMM -depth study on the basis of the existing research results , analyze the the HMM model humming feasibility and theoretical advantages , and its music search the improved model and algorithm design prototype music retrieval system . Melody signal Short time frame , the melodic feature vector is extracted , to examine these feature vector sequence in the statistical law , each melody were trained before the retrieval of modeling , and then calculate the probability of the test melody on these models , the selection of maximum probability as the search results to achieve effective retrieval of music . Specific work as follows : 1 . Analysis of the content-based music retrieval structure , given the common representation of audio information representation and MIDI music melody . Analysis of the the HMM model 's basic principle , outlined the three core issues of the model and the basic algorithm to solve three problems proposed HMM model to extract and identify the feasibility and theoretical advantages humming retrieval features . Analysis process of HTK toolkit training and to identify certain types of objects , as in HMM classifier , humming the music content classification to achieve the HMM classification based LFPC algorithm research . Discussed MID music feature extraction and selection , continuous hidden Markov model (CHMM) basis to create a melody model train and identify units of frames , and verified by experiments HMM model humming effective sex .

Related Dissertations

  1. Moving target trajectory based recognition system for human-computer interaction,TP391.41
  2. Research on Human Abnormal Behavior Detection and Recognition in Intelligent Video Surveillance,TP391.41
  3. Freeway incident detection based on video,TP391.41
  4. Investigation and Application of the Algorithm of Face Detection and Face Recognition,TP391.41
  5. Research and Implementation of Speaker Recognition System Based on VQ and HMM,TN912.34
  6. The Research and Application of Human Face Detection and Recognition Technology,TP391.41
  7. HMM-NRBF hybrid model of chaotic series modeling and application,TP183
  8. Chord Recognition with Beat Detection,TN912.34
  9. Hidden Markov Model Based Fingerprint Matching,TP391.41
  10. Research on Contourlet Transform and Application in Image Processing,TP391.41
  11. Researching of Mongolian Word Segment Based on Cascaded Hidden Markov Model,TP391.1
  12. Mandarin isolated word speech recognition technology research,TN912.34
  13. Key-dependent HMM -based musical chord recognition system,TN912.34
  14. A Research on Joint Source-Channel Decoding Algorithm of LDPC and Its Application,TN911.2
  15. K neighbors phased approach in the wireless LAN intrusion detection research,TN925.93
  16. The Chinese vowel length adjustment based speech recognition,TN912.34
  17. HMM and GMM based on natural and artificial blasting earthquake Recognition Algorithm,TN911.7
  18. Statistical Image Modeling and Image Segmentation,TP391.41
  19. Research and Implementation of Time Series Classification Based on Semi-supervised Learning,TP181
  20. Research on Human Motion Recognition Based on Video Sequences,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Retrieval machine
© 2012 www.DissertationTopic.Net  Mobile