carsayao / speech-recognition

Usage

Create /data at root and place wav files in /data in respective folders (i.e. bed wav commands go in /bed and /bed goes in /output), or change path in collect_features.py. Run python3 collect_features.py to generate /output folder. /output contains respective folders, each containing csv files of mid-term features. collect_features.py also generates short-term features which are saved in /output. Run python3 mean_features.py to generate a csv containing the mean of each folder.

Project Proposal

Team Members

• Carter Wilson
• Carlos Miguel Sayao
• Theodore Vixay
• Gabriel Miller

Objective

Visualize a given set of audio speech data using a combination SciPy, MatPlotLib, Tableau, and PyAudioAnalysis. Use clustering algorithms and techniques to group the audio data.

Approach

We will be using the dataset provided by the Kaggle Speech Recognition challenge. This includes 65,000 one-second long utterances of 30 short words by thousands of different people. Using amplitude as a measurement, we will process the data and use SciPy and MatPlotLib to provide basic waveform and spectrogram visualizations, with some aggregation. We will then use Tableau and PyAudioAnalysis to apply clustering techniques to see if we can find trends in visualizations of different words and types of words (verbs, consonants, etc.).

Team Structure

Our team consists of two sub-teams. One group works on classifying the dataset using machine learning techniques. Our group works on visualizing the dataset. The learning process will consist of regular meetings where we will study and work towards milestones.

Milestones

• Successful processing of the audio data
• Successful build of basic wave and spectrogram visualizations
• Aggregated visualizations of the audio data
• Build clustered visualizations of the audio data
• Find trends and interpret data

How to set up and use virtual env

pip3 install virtualenv

virtualenv -p python3 env

This will install virtualenv using pip3 and create a an env folder that will contain the libraries brought it.

source env/bin/activate

This will put your shell into the virtual environment where you will have the libraries brought it.

pip3 install -r requirements.txt

This will install the libraries that are found in the requirements.txt When you are done working, use deactivate to get out of the virtual environment