VSA and Voice Recognition Process
$3750-11250 MXN
Pagado a la entrega
VSA is defined as Voice Stress Analysis.
The programmer needs to have the understanding and knowledge of points of analysis and how to measure stress in a source of audio.
When a call with the customer ends an audio file gets generated by the system, the software needs to:
1. Recognize the number of voices in the audio file.
2. Separate the voices in two, generally the phone operator and the customer.
3. Perform VSA to each of the two categorized voices maybe splitting audio when silence or change of voice gets detected.
4. Summarize and as result its needed a metric of the call in quantitative variables, how much stress is in the call, who has the stress, the Operator or the Customer and when specifically saying determined words the stress was detected.
The use of voice recognition or voice triangulation, the use of voice transcription as CMUSphinx or other platforms may be required.
The software must be in Linux, for incorporating it to any kind of already made process starting with the gathering of the audio file in the filesystem, processing it and placing the results of each audio file in a SQL Database for the later processing of further action on the primary process funnel that can be a Call Center or other type of applications, a website, etc.
Nº del proyecto: #14663569
