KMITL

Permanent URI for this communityhttps://dspace.kmitl.ac.th/handle/kmitl/1

Browse

Settings

Access type: UnknownSubject: search.filters.subject.MFCC

Search Results

Now showing 1 - 4 of 4
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Improving the Sound Classification Accuracy Using CNN-LSTM and MFCC with Audio Augmentation for Diagnosing Respiratory Disease
    (2025-01-01)
    Phankokkruad, Manop
    ;
    Wacharawichanant, Sirirat
    Audio is vital information data for understanding various situations. A multitude of sound features can be explained by analysis through the audio signals. Numerous classification methods have been developed to study audio classification. This work studies the improvement of audio classification for the diagnosis of respiratory disease through the integration of audio data augmentation and CNN in conjunction with LSTM (CNN-LSTM). Furthermore, this paper focuses on audio data augmentation and feature extraction in the deep learning approach. This study proposed the CNN-LSTM model to diagnose respiratory disease by learning from the different audio datasets. The results reveal that the CNN-LSTM model attained an accuracy of 81.48%, precision of 0.8340, sensitivity of 0.6948, and F1-score of 0.7225. Considering the achieved F1-score, the CNN-LSTM model demonstrates a high level of diagnotic accuracy. Therefore, all evaluation evaluation parameters collectively indicate the robust performance of the proposed disease classification model.
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Segmentation of 24GHz RF Respiration Signal Using GMM
    (2019-11-01)
    Khaemphukhiao, Ratchaphon
    ;
    Phasukkit, Pattarapong
    This paper proposes a system for measuring rate of respiration of Newborn Phantom in a child incubator without contact with radar. In modern medicine, this non-contact system tends to replace the old systems that use contact measurement. The system of radar uses a 24 GHz continuous signal module, which has a region. that can detect breathing more clearly than low frequency. In addition, we have introduced the GMM technique to perform the signal segmentation for increasing the accuracy of the rate by compare with the contact signal to achieve 100% results.
      1
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    An Infant Cry Recognition based on Convolutional Neural Network Method
    (2019-11-01)
    Teeravajanadet, K.
    ;
    Siwilai, N.
    ;
    Thanaselanggul, K.
    ;
    Ponsiricharoenphan, N.
    ;
    Tungjitkusolmun, S.
    In this paper, an investigation of crying signal spectra is used to classify categories of infant cries. Three different types of crying considered in this work are hungry, sleepy and burping need. These cries are preprocessed and converted for calculation of Mel-Frequency Cepstral Coefficients (MFCC) before being classified by Convolutional Neural Network (CNN). Experimental results show that CNN based deep learning achieves high performance of 84%.
      1
  • Loading...
    Thumbnail Image
    Some of the metrics are blocked by your 
    consent settings
    Item type:Publication,
    Doppler radar for dynamic hand gesture recognition based on signal image processing
    (2019-07-01)
    Arthamanolap, Kongphum
    ;
    Gabbualoy, Somprasong
    ;
    Phasukkit, Pattarapong
    From previous researches, Doppler radar was used to detect signal for implement with many applications. Nevertheless, it is difficult to analyze for recognize object. At present, technique of deep learning in terms of signal processing and image processing are using in many researches to classify categories of data. In this paper, signal image was used by deep learning to classify hand gesture by receiving signals from 24GHz transceiver: BGT24MTR11. We transformed the signals to images for 3 categories including Spectrogram, Time domain from original signal and feature MFCC graph. After that those of converted image will be trained by Deep learning for classify the hand gesture types. From the result of this experiment has been shown that signal image can be used to recognize hand gesture and Spectrogram graph makes the highest accuracy as 94%.