ePrints.FRI - University of Ljubljana, Faculty of Computer and Information Science

System for audio capture and classification of baby cry samples

Saša Saftić (2017) System for audio capture and classification of baby cry samples. MSc thesis.

Download (2903Kb)


    We explore multiclass classification of infants' cries and the relation between the age of the infant and the accuracy of classification. Additionally we explore secure cloud storage and cloud data processing. We compare several state-of-the-art multiclass classification models with recurrent neural networks. Classification accuracy was obtained on data from infants of various ages. For data storage and processing we used the Django Rest API and the opensource cloud platform OpenStack. Multiclass classification models successfully differentiated between different classes of crying, but no age effect has been found. We have demonstrated the aptness of the Django Rest API and OpenStack platform for data storing and processing in the cloud.

    Item Type: Thesis (MSc thesis)
    Keywords: infant cry, data acquisition, audio analysis, feature extraction, classification, cloud processing, cloud security
    Number of Pages: 80
    Language of Content: Slovenian
    Mentor / Comentors:
    Name and SurnameIDFunction
    izr. prof. dr. Mojca Ciglarič256Mentor
    prof. dr. Blaž Zupan106Comentor
    Link to COBISS: http://www.cobiss.si/scripts/cobiss?command=search&base=51012&select=(ID=1537412803)
    Institution: University of Ljubljana
    Department: Faculty of Computer and Information Science
    Item ID: 3785
    Date Deposited: 10 Feb 2017 13:16
    Last Modified: 21 Apr 2017 10:35
    URI: http://eprints.fri.uni-lj.si/id/eprint/3785

    Actions (login required)

    View Item