Methods for segmentation and classification of swallowing instants from the feeding sound of newborn infants

Koyuncu, Abdullah Onur
Statistics such as swallow frequency, the average time between rhythmic swallows and the maximum number of rhythmic swallows can be related to the feeding maturity of infants. Therefore, detecting swallow segments automatically from an acoustical feeding signal can be considered as a decision support mechanism for neonatologists. This thesis includes different approaches for the analysis of infant's feeding sounds and proposes two different pattern recognition methodologies, segmentation followed by classification and classification followed by merging, for auto-segmentation and classification of swallowing instants. Data from 52 infant subjects are used, in which acoustic feeding signals are recorded with a digital stethoscope. Multiple learning algorithms such as Gaussian mixture models (GMM), support vector machines (SVM) and hidden Markov models (HMM) are used to discriminate swallowing sounds from other sound activities. A comprehensive set of feature extraction methods in time and frequency domain are investigated for the representation of captured acoustic signals. Moreover, feature selection methods are examined thoroughly to improve the representation power of feature vectors. Experimental comparison in terms of precision, recall and F1 scores of eight different paths to segment and classify swallow instants is made. The results show that the first approach segments the swallow episodes with lower performance as the error in the segmentation also affects the classification performance negatively. On the other hand, best results are obtained in the second approach where binary and 3 class SVM classifiers are applied with purpose-specific finite state machine algorithms. In the time duration based performance evaluation, the F1 scores are obtained as almost equal to 0.70 for both methods. On the other hand, they are computed as nearly 0.81 in the event based one.


Comparison of three different primer sets for sexing birds
Çakmak, Emel; Peksen, Cigdem Akin; Bilgin, Cemal Can (2017-01-01)
Because many bird species are monomorphic or only sexually dimorphic in adult stages, it is difficult to determine their sexes, which may cause significant problems in population and conservation studies. DNA-based sexing relies on the chromodomain helicase DNA binding (CHD) gene located on the W chromosome and its homolog on the Z chromosome, giving distinct banding patterns on agarose gel as a result of length differences in intronic regions within this gene. We used 3 specific primer sets, CHD1F/CHD1R, 2...
Breeding biology, population size and spatial distribution of a common nightingale (Luscinia megarhynchos Brehm, 1831) Population at Yalıncak (Ankara)
Özbahar, İlker; Bilgin, Cemal Can; Department of Biology (2005)
Many bird populations are known to show strong territoriality as well as high site fidelity during breeding. The Common Nightingale (Luscinia megarhynchos) is one of many such migratory species that annually occupies the same favourable habitat. Especially dominant males prefer to breed in the same area every year, and high fidelity probably assures high breeding success. This study aims to investigate breeding biology, population size and spatial distribution of nightingales at a small area in Yalıncak wit...
Optimization of Frequency Lowering Algorithms for Getting the Highest Speech Intelligibility Improvement by Hearing Loss Simulation
ARIÖZ, UMUT; Günel Kılıç, Banu (2015-06-01)
High frequency hearing loss is a growing problem for both children and adults. To overcome this impairment, different frequency lowering methods (FLMs) were tried from 1930s, however no satisfaction was provided up to now. In this study, for getting higher speech intelligibility, eight combinations of FLMs which were designed originally were tried with simulated sounds onto normal hearing subjects. These improvements were calculated by the difference with standard hearing aid method, amplification. High fre...
Narrowing Perceptual Sensitivity to the Native Language in Infancy: Exogenous Influences on Developmental Timing
Elsabbagh , Mayada; Hohenberger, Annette; Campos, Ruth; Van Herwegen, Jo; Serres, Josette; de Schonen, Scania; Aschersleben, Gisa; Karmiloff-Smith, Annette (MDPI AG, 2013-2-6)
The infancy literature situates the perceptual narrowing of speech sounds at around 10 months of age, but little is known about the mechanisms that influence individual differences in this developmental milestone. We hypothesized that such differences might in part be explained by characteristics of mother-child interaction. Infant sensitivity to syllables from their native tongue was compared longitudinally to sensitivity to non-native phonemes, at 6 months and again at 10 months. We replicated previous fi...
Baykal, Nazife; YALABIK, N; Erkmen, Aydan Müşerref (1994-01-01)
Doppler umbilical artery blood flow velocity waveform measurement is used in perinatal surveillance for the evaluation of pregnancy status. There is an ongoing debate on the predictive value of doppler measurements concerning the critical effect of the selection of parameters for the evaluation of doppler output. In this paper, we describe how neural network methods can be used both to discover relevant classification features and subsequently to classify patients. Classification accuracy varied from 92-...
Citation Formats
A. O. Koyuncu, “Methods for segmentation and classification of swallowing instants from the feeding sound of newborn infants,” Thesis (M.S.) -- Graduate School of Natural and Applied Sciences. Electrical and Electronics Engineering., Middle East Technical University, 2019.