June 2019
Intermediate to advanced
308 pages
7h 21m
English
It is essential to explore a dataset before applying DL algorithms on the data. To explore, firstly, we can run the audio signal (.wav) to the image converter, wav2image.py (available in Chapter 4, Audio/Speech/Voice Recognition in IoT code directory), to see how the spectrum image looks. This will produce images, as shown. The following screenshot shows converted images for an on command:

The following screenshot shows converted images for an off command. As we can see from the screenshots, their color distributions are different, which will be exploited by the DL algorithms in order to recognize them:
We can also carry out ...
Read now
Unlock full access