peruginia/devicehive-audio-analysis

Devicehive Audio Analysis

Audio classification feature demo
Detailed description can be found here

Installation

• Get a copy of this repo
• Install system packages

sudo apt-get install libportaudio2 portaudio19-dev

• Install python requirements

pip install -r requirements.txt

• Download and extract saved models to source directory

wget https://s3.amazonaws.com/audioanalysis/models.tar.gz
tar -xzf models.tar.gz

Running

To process prerecorded wav file

run

python parse_file.py path_to_your_file.wav

Note: file should have 16000 rate

To capture and process audio from mic

run

python capture.py

It will capture and process samples in a loop.
To get info about parameters run

python capture.py --help

To start web server

run

python daemon.py

By default you can reach it on http://127.0.0.1:8000
It will:

• Capture data form your mic
• Process data
• Send predictions to web interface
• Send predictions to devicehive

Also you can configure your devicehive connection though this web interface.

Useful info

To train classification model next resources have been used:

• Google AudioSet
• YouTube-8M model
• Tensorflow vggish model

You can try to train model with more steps/samples to get more accuracy.