Dynamics and Periodicity Based Multirate Fast Transient-Sound Detection. Yang, J. & Hilmes, P. In 2018 26th European Signal Processing Conference (EUSIPCO), pages 2449-2453, Sep., 2018.
Dynamics and Periodicity Based Multirate Fast Transient-Sound Detection [pdf]Paper  doi  abstract   bibtex   
This paper proposes an efficient real-time multirate fast transient-sound detection algorithm on the basis of emerging microphone array configuration intended for multimedia signal processing application systems such as digital smart home. The proposed detection algorithm first extracts the dynamics and periodicity features, then trains the model parameters of these features on Amazon machine learning platform. The real-time testing results have shown that the proposed algorithm with the trained model parameters can not only achieve the optimum detection performance in all various noisy conditions but also reject all kinds of interferences including undesired voice and other unrelated transient-sounds. In comparison with the existing algorithms, the proposed detection algorithm significantly improves the false negative and false positive performance. In addition, the proposed multirate strategy dramatically reduces the computational complexity and processing latency so that the proposed algorithm can serve as a much more practical solution for the digital smart home related applications.
@InProceedings{8553506,
  author = {J. Yang and P. Hilmes},
  booktitle = {2018 26th European Signal Processing Conference (EUSIPCO)},
  title = {Dynamics and Periodicity Based Multirate Fast Transient-Sound Detection},
  year = {2018},
  pages = {2449-2453},
  abstract = {This paper proposes an efficient real-time multirate fast transient-sound detection algorithm on the basis of emerging microphone array configuration intended for multimedia signal processing application systems such as digital smart home. The proposed detection algorithm first extracts the dynamics and periodicity features, then trains the model parameters of these features on Amazon machine learning platform. The real-time testing results have shown that the proposed algorithm with the trained model parameters can not only achieve the optimum detection performance in all various noisy conditions but also reject all kinds of interferences including undesired voice and other unrelated transient-sounds. In comparison with the existing algorithms, the proposed detection algorithm significantly improves the false negative and false positive performance. In addition, the proposed multirate strategy dramatically reduces the computational complexity and processing latency so that the proposed algorithm can serve as a much more practical solution for the digital smart home related applications.},
  keywords = {acoustic signal processing;filtering theory;home automation;learning (artificial intelligence);microphone arrays;signal detection;microphone array configuration;multimedia signal;periodicity features;Amazon machine learning platform;optimum detection performance;false negative performance;false positive performance;multirate strategy;digital smart home related applications;realtime multirate fast transient-sound detection algorithm;Signal processing algorithms;Microphones;Heuristic algorithms;Feature extraction;Signal processing;Smart homes;Transient analysis;feature extraction;fast transient-sound detection;sound source localization;digital-positioning system;smart home},
  doi = {10.23919/EUSIPCO.2018.8553506},
  issn = {2076-1465},
  month = {Sep.},
  url = {https://www.eurasip.org/proceedings/eusipco/eusipco2018/papers/1570435103.pdf},
}
Downloads: 0