Detecting signals from .wav files after deep learning network is trained

1 次查看（过去 30 天）

Ashley Cook 2021-6-7

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/850185-detecting-signals-from-wav-files-after-deep-learning-network-is-trained

评论： ytzhak goussha 2021-6-30

Are there any follow up tutorials or examples on how to detect sounds in .wav files after you train a network to classify spectrograms? The 'deep learning with MATLAB course' teaches how to classify spectrograms, but in order for this to be useful in my work I wanted to be able to then use the the trained network to evaluate ~30 min .wav files and detect a desired sound type. Any links or suggestions are welcome.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

ytzhak goussha 2021-6-30

Hey,

reshape your vector into a matrix, it possible that you might have to zero pad it or truncate it to able to reashe the vector into a matrix with desider number or row.
use audioFeatureExtractor to create an extractor with desird parameters and use it to extract spectrograms from you audio matrix
feed the spectrograms to your model. if your spectrograms are too large, feed the matrix in small batches

请先登录，再进行评论。

请先登录，再回答此问题。