Background

In recent years, cybercriminals have increasingly exploited VoIP technology and caller ID spoofing to conduct frauds, extortion, phishing scams, and fake digital arrests. These calls often appear legitimate, impersonating government agencies, banks, or relatives. Traditional telecom infrastructures and mobile OSs struggle to effectively detect and block such calls in real-time, especially when they are routed through international VoIP gateways or masked using spoofing tools.

India and other countries are witnessing a surge in such fraudulent calls, which often go unreported or undetected until financial or emotional harm has occurred. There is an urgent need for a technological solution that can detect, classify, and alert users about malicious voice calls before or during the call.

Problem Statement

There is no accessible, real-time system to detect and classify voice calls — particularly those made using VoIP networks, spoofed caller IDs, or known spam sources —especially in resource-constrained mobile environments.

Key Objectives

Design and build a Real-Time Call Analysis System (mobile app or backend platform) that can:

• Improve public safety by protecting against fraudulent or harmful voice calls
• Enable quick response and awareness during or before the call
• Empower users and law enforcement with intelligent call classification

Functional Requirements

1.  Identify and flag incoming calls that are:

o Spoofed (Caller ID mismatch, falsified metadata)

o Spam or robocalls (Telemarketing, mass-calling bots)

o VoIP-based fraud calls (Overseas calls using disguised SIP protocols)

2.  Analyse calls in real time or near-real time using techniques such as:

o SIP/VoIP header analysis

o Caller ID validation via network APIs

o Historical user behaviour (e.g., frequency, duration, timing)

o Crowd-sourced reports and spam number databases

o Acoustic fingerprinting (optional advanced feature)

3.  Alert users immediately with a visual or audio cue:

o “Suspicious Call: Possible Spoof"

o “Spam Likely: High Risk VoIP Number"

o “Verified Call: Safe"

4.  Optional Features:

o Allow users to report a call as fraud/spam/spoof

o Generate call metadata report for law enforcement

o Integrate with Do Not Disturb (DND) and TRAI databases

o Provide statistics and heatmaps for scam call trends

Evaluation Criteria

• Real-time detection accuracy and minimal false positives
• Use of intelligent detection (heuristics, signatures, ML, etc.)
• Performance under low-resource (mobile) environments
• UI/UX design for alerts and feedback loop
• Privacy preservation and data compliance

Bonus Points

• Working mobile app (Android/iOS)
• Use of live telecom data or simulation of SIP call metadata
• Support for multiple languages or regional awareness
• Demonstration with sample spoof/spam call scenarios