# 🚀 AI-Powered Forensic Scoring Engine (Phase 3.4)

[KhairnarLokesh]

📌 Title

Implement AI-Powered Forensic Soundness Scoring Engine

---

🎯 Goal

Create a backend service that calculates a Forensic Soundness Score (0–100) for digital evidence based on integrity, metadata completeness, and chain-of-custody verification.

The score will help analysts quickly assess the reliability and admissibility strength of uploaded digital evidence.

---

🧠 Scoring Logic (Total = 100 Points)

The scoring engine should assign points based on the following criteria:

Criteria	Points
✅ Blockchain verification status	+40
✅ IPFS Content ID (CID) present	+20
✅ Metadata completeness (Location, GPS, Timestamp present)	+20
✅ Chain of custody history exists	+20
Maximum Score	100

---

⚙️ Functional Requirements

1️⃣ Backend Service

File to Create:

services/forensicScoreService.js

Responsibilities:

• Accept evidence object as input
• Evaluate all scoring parameters
• Return:
  • Final score (0–100)
  • Breakdown of points per category
  • Score classification (Optional enhancement)

Example Output Format:

{
  "score": 85,
  "breakdown": {
    "blockchain": 40,
    "ipfs": 20,
    "metadata": 15,
    "chainOfCustody": 10
  },
  "classification": "High Integrity"
}

---

2️⃣ Scoring Rules Detail

🔐 Blockchain Verification (+40)

• If blockchain hash exists and verified → +40
• If exists but not verified → +20 (optional improvement)
• If not present → 0

📦 IPFS CID Presence (+20)

• If CID exists → +20
• Else → 0

📍 Metadata Completeness (+20)

Check if all required fields are present:

• Location
• GPS Coordinates
• Timestamp

Scoring:

• All 3 present → +20
• 2 present → +15
• 1 present → +10
• None → 0

🔗 Chain of Custody (+20)

• If custody history length ≥ 1 → +20
• If partial/incomplete → +10 (optional improvement)
• If none → 0

---

🖥️ Frontend Integration

File to Modify:

public/dashboard-analyst.html

Add:

• Display Forensic Soundness Score (Prominent UI)
• Show:
  • Numeric score
  • Visual indicator (progress bar / circular gauge)
  • Score breakdown
  • Classification badge (Low / Medium / High Integrity)

Suggested Classification Ranges:

Score Range	Label
80–100	🟢 High Integrity
50–79	🟡 Medium Integrity
0–49	🔴 Low Integrity

---

🔁 API Integration (If Required)

If evidence is fetched via API:

• Ensure score is calculated before sending response
  OR
• Create endpoint:

GET /api/evidence/:id/forensic-score

---

🧪 Testing Requirements

• Test with:
  • Fully verified evidence (Expect 100)
  • Partial metadata
  • No blockchain verification
  • Missing CID
• Validate score accuracy
• Edge case: Empty evidence object

---

📊 Acceptance Criteria

• forensicScoreService.js implemented
• Scoring logic correctly calculates 0–100
• Returns structured breakdown
• Dashboard displays score clearly
• Score updates dynamically when evidence changes
• Tested with multiple evidence states

---

💡 Future Enhancements (Optional)

• Machine Learning-based anomaly detection
• Weight adjustment via admin panel
• Risk prediction model
• Auto-flag low score evidence

---

🏁 Expected Outcome

A fully functional AI-powered forensic scoring engine that:

• Quantifies evidence reliability
• Improves analyst decision-making
• Strengthens legal defensibility
• Enhances transparency of digital chain-of-custody

---

Priority: High
Phase: 3.4
Type: Feature Implementation