In previous studies, general approach that is being
followed in most daily sound recognition systems consists of
three stages: sound acquisition, feature extraction and
classification. Various researchers have worked on data
collection of daily sounds using different sensors which can
be used for sound classification e.g., [1], [3], [4]. In [2],
authors have used a combination of sound and vibration on
the floor for detection of fall situations inside home. In [5],
the author has used a combination of two microphones for
fall detection recordings. Features play a very important part
in the classification task. They provide a reduced numerical
representation of the sound. If features used are not able to
characterize sounds in different classes, then even an
accurate classifier will give bad results. Research work in [6],
In previous studies, general approach that is beingfollowed in most daily sound recognition systems consists ofthree stages: sound acquisition, feature extraction andclassification. Various researchers have worked on datacollection of daily sounds using different sensors which canbe used for sound classification e.g., [1], [3], [4]. In [2],authors have used a combination of sound and vibration onthe floor for detection of fall situations inside home. In [5],the author has used a combination of two microphones forfall detection recordings. Features play a very important partin the classification task. They provide a reduced numericalrepresentation of the sound. If features used are not able tocharacterize sounds in different classes, then even anaccurate classifier will give bad results. Research work in [6],
การแปล กรุณารอสักครู่..