ZONN.ai Forensic Report

Case · FF983105 · IMAGE

Analyzed byAnonymous
ZONN Analysis
0

Inconclusive

Signals are mixed or weak. We can't tell with confidence — context, source, and your own judgement matter here.

Signal ConfidenceWeak · 18/100

Analysed Specimen

Analysed content
No flagged regions

Heads up — 4 things to know

Why this analysis might be off

We highlight every disagreement and unusual signal we found so you can judge for yourself. Stronger warnings come first; informational notes are at the bottom.

Multiple detectors unreachable

Note

4 ML detectors did not respond (Ml Img Cifake, Ml Img Deepfake, Ml Img 3class, Ml Python Siglip). The verdict was computed with reduced evidence; reliability is lower than usual.

Signals do not converge

Note

Score 51/100 lands in the middle of the inconclusive zone. Detectors did not agree on a story — review multiple lines of evidence below.

Weak overall confidence

Note

Aggregate verdict confidence is 18/100. Several detectors returned uncertain answers or were offline. Read the verdict as a guideline, not as a final answer.

No ICC color profile

AI evidence

The image does not embed an ICC color profile. Real camera files almost always carry sRGB or Adobe RGB profiles — a missing profile is often a sign of generator output or re-encoded media.

Why this verdict

  • Error Level Analysisread real · 12/100

    Error Level Analysis. Re-saves and diffs to expose uneven compression regions.

  • Color Distributionread real · 31/100

    Global color distribution shape vs natural-photo baselines.

Model Agreement

100%

Variance across 8 ML detectors. Higher agreement means the models converged on the same reading; lower agreement means treat the verdict with care.

Evidence — 18 detectors reviewed

What each detector saw

Each detector independently gave this imagea score from 0 (definitely real) to 100 (definitely AI). The score above is their weighted consensus — detectors with higher confidence count more. No single detector decides; you read the spread.

ML Models8 detectors · mean 50
▸ expand
AI Image Detector
50
Deepfake Detector
50
3-Class Detector
50
SigLIP AI Detector
50
RIGID (DINOv2)
50
Fourier Zero-Shot
50
NPR Artifacts
50
GramNet Texture
50
Pixel & Frequency Forensics7 detectors · mean 40
▸ expand
Error Level Analysis
12
Color Distribution
31
Pixel Analysis
35
Noise Pattern
62
Edge Consistency
43
Frequency Analysis
44
Compression Quality
55
Provenance & Metadata3 detectors · mean 57
▸ expand
ICC Profile
62
Metadata
59
C2PA Provenance
50

Image Quality

Dimensions
1888 × 2228 px
Aspect
0.847
File size
7.11 MB
Bytes / pixel
1.772

Edge Consistency

CV 0.713
Cell 1: 1.5955Cell 2: 11.4376Cell 3: 11.3808Cell 4: 1.3141Cell 5: 1.5002Cell 6: 14.9742Cell 7: 15.8920Cell 8: 1.8151Cell 9: 4.9359Cell 10: 31.9184Cell 11: 15.8914Cell 12: 8.9757Cell 13: 13.9662Cell 14: 25.9256Cell 15: 20.5115Cell 16: 12.6127

Per-region edge density (4 × 4 grid). Uneven distribution may indicate localized editing or splicing; uniform fields are typical of fully synthetic outputs.

Range: 1.314131.9184

Noise Fingerprint

Variance
102.36
Std deviation
10.12
Mean
-0.0
Spatial corr.
2.935
Mean Δ
1.71
σ
2.27
CV
1.327
Uniformity
-0.327

Provenance

Source Dossier

PlatformDirect upload
Author
Content Typeimage
Analyzed OnMay 5, 2026, 9:34 PM
Analyzed byAnonymous