An expert at the Abubakar Tafawa Balewa University, Kano, in northern Nigeria, Engr Mahmood Abdulhameed has designed and constructed a customer voice identification system for the Nigerian telecommunications industry that would enable the telecos and other intelligence agencies search criminal by their voices over telecommunication system.
The project, which was the result of a grant from the Nigerian Communications Commission (NCC), revealed that the telecos now have a system that would automatically identify the voice of a customer even among other voices using the speaker identification model.
According to Abdulhameed, in the research report sighted by TechnologyMirror the system would permit Telecom industries to search for people by their voices after making calls or voice chats, adding that there is no need to know customer number or internet protocol (IP) address and that it only needs to have his voice saved in the system.
“It works based on the idea that if I knew your voice then, phone number and IP addresses are not required once customer voice is in the system, the system will automatically check if the voice made any call or voice chat and once detected all the calls he made with corresponding dates, phone number used and from which location he made any of the calls can be traced by the telecommunication company including tower used to originate the call”, he said.
While stressing the need for the system, he said:”the major concern in telecommunication industries due to the fact that we have a lot of security challenges like Boko Haram, Niger Delta Militant, Armed Robbers, Hackers,Corruption, and Kidnapers etc. Pictures of all known wanted Boko Haram are available and most of their voices is also available like Shekau.”
He disclosed that human voice contains large amount of information about a speaker such as his emotions,
identity, language as well as the message communicated saying that the database for the project was initially an open source Database which contains the Census (AN4) database audio files and the design was tested on that Database but Based on Request by NCC that the Database must contain our Nigeria Local speaker at least male and
female Hausa, Yoruba and Igbo.
“At the end of this research the aim of the project was achieved because speaker identification was developed and the system can be used for customer identification by their voices in telecommunication industry with higher accuracy”, he stated.
He recommended that NCC should be able to test this project on telecommunication database and on larger database. adding that it should check the limitation of the project “based on telecommunication data and finally we recommended that for larger database, higher specification Computer should be used especially in term of processor speed.”