/Listenin

A standalone device that detects depression through vocal input.

Primary LanguageJupyter Notebook

Listenin

Overview

This project aims to propose a device that takes human voice as input to diagnose depression. It would analyse the spectrogram of the person’s voice, run it through a Machine Learning model (Convolutional Neural Network) generated for this purpose and produce the output of whether the person is depressed or not.

Methodology

The project is carried out in the following steps:

asdf

qwert

Application

This device and system created can directly be used by people to assess their mental health. General medical practitioners can use the standalone device to screen for clinical depression in rural health camps. It can also be integrated with smart home assistants like Google home or Alexa which take in voice input.

Advantages

Existing applications in playstore are based on text and not on voice. Diagnosis with voice has proved to be more accurate. The model we propose uses the vocal parameters and not the content, hence there is no issues of infringement of privacy.

zxcv