Real Time Speech Emotion Recognition and Voice Analytics in Jails

  • Home
  • Real Time Speech Emotion Recognition and Voice Analytics in Jails
REAL TIME SPEECH EMOTION RECOGNITION AND VOICE ANALYTICS IN JAILS

Real Time Speech Emotion Recognition and Voice Analytics in Jails

Category: Case Studies

The emotional well-being of inmates is of great importance to correctional officers. Correctional institutions are often plagued by incident of violence between inmates, between inmates and correctional officers. There are also numerous disciplinary incidents involving inmates being disruptive and not following instructions, thereby threatening the safety of the officers and other inmates. It is understood that inmates under emotional stress act out against other inmates or the jail / prison staff.  It has also been observed that inmates’ emotional states change, sometimes for the better and sometimes for the worse, following telephone interactions with their family members. Sometimes, these interactions result in incidents of violence and in some cases, even in attempts at suicide.

Correctional officers are often trained to be observant of the emotional state of the inmates. They observe inmates looking for emotional cues when individuals behave in unusual manner. This technique relies on the experience level and the training received by the officers. Being able to monitor the emotional state of the inmates, in an automated manner, during these exchanges with their families / friends is of great value to the correctional facilities.

KTM – Speech Emotion Recognition (SER), applies Digital Signal Processing, DSP, to the audio signals to extract various details about the conversation, then applies Deep Learning to predict the emotional state of the speakers. The algorithms work in a real-time manner and are able to discern the changes in emotional state of the speakers. Once the emotional features are detected, appropriate alerts are generated.

When combined with appropriate inmate telephone systems and alerting systems, KTM – SER:

  1. Helps monitor the emotional state of inmates using voice analytics.
  2. Flags potential emotional breakdowns, thereby permitting the prison officials to intervene ahead of time if needed.
  3. Helps mitigate suicide risks.
  4. Helps prevent violent outbreaks following emotional trauma