ZONN.ai Forensic Report

Case · ACC70C3C · IMAGE

Analyzed byAnonymous
ZONN Analysis
0

Probably AI-generated

More signals lean toward AI generation than real, but some give weaker readings. Treat with caution.

Signal ConfidenceWeak · 17/100

Analysed Specimen

Analysed content
No flagged regions

Heads up — 3 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 Deepfake, Ml Img Cifake, Ml Img 3class, Ml Python Siglip). The verdict was computed with reduced evidence; reliability is lower than usual.

Weak overall confidence

Note

Aggregate verdict confidence is 17/100. 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.

Origin Check

Trace this image elsewhere

Cross-reference the source against major reverse-image services. Each link opens in a new tab with the image URL preloaded — ZONN.ai does not re-upload the image.

Why this verdict

  • Error Level Analysisread real · 24/100

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

  • Pixel Analysisflagged AI · 75/100

    Low-level pixel statistics: saturation cliffs, channel histograms, banding.

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
Deepfake Detector
50
AI Image 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 47
▸ expand
Error Level Analysis
24
Pixel Analysis
75
Frequency Analysis
28
Noise Pattern
65
Color Distribution
40
Compression Quality
45
Edge Consistency
51
Provenance & Metadata3 detectors · mean 56
▸ expand
ICC Profile
62
Metadata
55
C2PA Provenance
50

Image Quality

Dimensions
1037 × 1516 px
Aspect
0.684
File size
225.1 KB
Bytes / pixel
0.147

Edge Consistency

CV 0.478
Cell 1: 8.1059Cell 2: 18.0257Cell 3: 10.1962Cell 4: 6.4539Cell 5: 14.1348Cell 6: 27.6013Cell 7: 18.3578Cell 8: 9.9626Cell 9: 7.0013Cell 10: 27.8053Cell 11: 16.4633Cell 12: 6.0748Cell 13: 9.2365Cell 14: 13.6511Cell 15: 21.7018Cell 16: 12.1129

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

Range: 6.074827.8053

Noise Fingerprint

Variance
84.04
Std deviation
9.17
Mean
-0.0
Spatial corr.
3.248
Mean Δ
2.15
σ
2.45
CV
1.136
Uniformity
-0.136

Provenance

Source Dossier

PlatformDirect upload
Author
Content Typeimage
Analyzed OnMay 6, 2026, 10:54 PM
Analyzed byAnonymous