AMMON : A Speech Analysis Library for Analyzing Affect , Stress , and Mental Health on Mobile Phones

The human voice encodes a wealth of information about emotion, mood and mental state. With mobile phones this information is potentially available to a host of applications. In this paper we describe the AMMON (Affective and Mentalhealth MONitor) library, a low footprint C library designed for widely available phones. The library incorporates both core features for emotion recognition (from the Interspeech 2009 emotion recognition challenge), and the most important features for mental health analysis (glottal timing features). To comfortably run the library on feature phones (the most widely-used class of phones today), we implemented most of the routines in fixed-point arithmetic, and minimized computational and memory footprint. While there are still floating-point routines to be revised in fixed-point, on identical test data, emotion and mental stress classification accuracy was indistinguishable from a state-of-the-art reference system running on a PC.