Timing based source separation

dc.contributor.advisorHøst-Madsen, Anders
dc.contributor.authorYoung, Jeremy Brian Wai Mun
dc.contributor.departmentElectrical Engineering
dc.date.accessioned2020-02-20T18:12:28Z
dc.date.available2020-02-20T18:12:28Z
dc.date.issued2019
dc.description.abstractThe motivation of this work is to use statistical signal processing to help to separate mixtures of marine mammal vocalizations, in particular sperm whale click trains. It is observed that clicks from a single whale are spaced at regular intervals which we exploit to do the separation. To begin, we consider an idealized problem: each source emits at impulses at iid intervals according to some distribution. That is, the impulse times for each source constitute a renewal process. These timing distributions induce a likelihood function for the impulse times given some assignment. We present an algorithm inspired by the Viterbi algorithm to give assignments that maximizes this likelihood. Additionally, we provide more computationally feasible approximations of this algorithm. We verify these algorithms with a derived lower bound. Furthermore, we provide methods to estimate the timing distribution parameters and the total number of sources using alternating maximization and the minimum description length. We also attempted to use timing to help improve detection methods for impulses using a Bayesian framework. Unfortunately, the results were impractical, but the concepts were used to generate a classification using impulse shape information timing information. Finally we explore modifications to make the algorithm work on actual sperm whale click mixtures.
dc.description.degreePh.D.
dc.identifier.urihttp://hdl.handle.net/10125/66273
dc.languageeng
dc.publisherUniversity of Hawaii at Manoa
dc.subjectSource separation (Signal processing)
dc.subjectSignal processing--Mathematics
dc.subjectSignal processing--Statistical methods
dc.subjectSperm whale
dc.subjectSound production by animals
dc.titleTiming based source separation
dc.typeThesis
dc.type.dcmiText
local.identifier.alturihttp://dissertations.umi.com/hawii:10462

Files

Original bundle
Now showing 1 - 2 of 2
Loading...
Thumbnail Image
Name:
Young_hawii_0085A_10462.pdf
Size:
4.47 MB
Format:
Adobe Portable Document Format
No Thumbnail Available
Name:
code_contact.txt
Size:
124 B
Format:
Plain Text