Remote PTSD monitoring and diagnosis using an automated system

In this proposal, Intelligent Automation, Inc., propose an automated system that remotely monitors and tracks the patient’s mental healthiness. The proposed system first extract acoustic features such as "F0-amplitude", "F0-6db-bandwidth", Speaking rate, Percent pause time and Vocalization/pause ratio from patient'' voice. Then the patients’ voice will also be separated into phonemes and log likelihood scores will be generated by matching these phonemes with their Gaussian models. These log likelihood scores will be incorporated as part of feature sets. An advanced feature selection method, Piecewise Linear Orthogonal Floating Search (PLOFS) will be used to identify, select and map the most relevant features to the patient’s mental healthiness. Finally, the selected features will then pass through a novel committee machine-based regression algorithm to estimate the patient’s mental health “score”. The proposed system monitors each individual’s mental healthiness in real time and even tracks his/her mental health by refining the general model into his/her own personal model.