Voice traffic watermarking method for cellular-to-VoIP calls

Abstract

In recent years, voice phishing scams have caused significant damage, using cellular-to-VoIP calls such as SIM box calls and CMC calls. Criminals use cellular-to-VoIP calls to disguise international internet calls as domestic calls. The equipment supporting cellular-to-VoIP is often located in hard-to-detect places, and even if detected, direct information about overseas criminals is difficult to obtain, leading criminals to install new equipment to continue their activities.

This paper aims to assist in the investigation of voice phishing crimes by analyzing the call protocols of cellular-to-VoIP calls and proposing a method to apply watermarking technology to voice traffic. Specifically, we understand the protocol characteristics of SIM box calls and CMC calls and explain why existing voice traffic watermarking methods do not apply to these types of calls. Watermarking for cellular-to-VoIP call traffic needs to be robust, as the transition from cellular to VoIP occurs. It also requires direct insertion from the device and must remain inaudible to the recipient. This paper develops a watermark that meets these requirements and demonstrates its distinguishability from typical call traffic.