Sociofeedback by google glass

Abstract

This project concentrates on the development of an Android application for Sociofeedback. The available real-time Sociofeedback system and applications requires a series of devices to analyse the audio input and perform classification. In this project, all the signal processing and classification are aimed to be accomplished on the mobile devices themselves. Two platforms are discussed in this report — Google Glass and Android mobile phone. The report illustrates the development process with reference to the system development life cycle (SDLC). The functionalities of the applications have been varying through the weekly meetings. This application initially takes monologues as audio input to identify the speech mannerisms and provide feedback in order to enhance users’ presentation skills. Thinking of distributing such application to broader audience, we start to accommodate the application for two-person. The Sociofeedback app can further fit into workplace such as call centres and police offices. Users are able to select their preferred low-level features to be monitored, including volume, pitch, speaking percentage, and MFCC so far. The selected low-level features will be shown graphically on the app. Twelves low-level features are derived from the volume, pitch and MFCC of the speech from the two users to perform classification. Google Glass is one of the most cutting-edge wirable devices in the market. It is light, unsophisticated and portable. Although the processing capability of Google Glass is not satisfying in this project, the future generation of such smart glass would be an appreciable platform for this app. We use Android phone as an alternative device to showcase the application. The iterative design procedure will be elaborated in this report.