ZONN.ai Forensic Report

Case · 28FF79C6 · IMAGE

Analyzed byAnonymous
ZONN Analysis
0

Probably real

More signals lean toward real than AI, but some give weaker readings. Worth a second look on close inspection.

Signal ConfidenceWeak · 25/100

Analysed Specimen

Original analysed image
Forensic suspicion heatmap
OriginalHeatmap
POS55/100
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

3 ML detectors did not respond (Ml Python Siglip, Ml Commfor, Itsnotai V2). The verdict was computed with reduced evidence; reliability is lower than usual.

Upsampling artifacts in the frequency domain

AI evidence

FFT analysis found strong upsampling patterns — a fingerprint of diffusion-model VAE decoders (latent → pixel-space upscale).

Weak overall confidence

Note

Aggregate verdict confidence is 25/100. Several detectors returned uncertain answers or were offline. Read the verdict as a guideline, not as a final answer.

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.

Google Lens logoGoogle LensYandex logoYandexTinEye logoTinEyeBing logoBing Visual

Why this verdict

  • Noise Patternread real · 25/100

    Sensor-noise (PRNU) fingerprint check. Real cameras leave camera-unique noise.

  • Frequency Analysisflagged AI · 73/100

    Fast-Fourier inspection. Diffusion outputs leave periodic spectral peaks.

Model Agreement

100%

Variance across 4 ML detectors. Higher agreement means the models converged on the same reading; lower agreement means treat the verdict with care.

Evidence — 14 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 Models4 detectors · mean 50
▸ expand
SigLIP AI Detector
50
CommFor (4803 Generators)
50
INA v2 (FLUX/MJ)
50
Manipulation Map (IML-ViT)
50
Pixel & Frequency Forensics7 detectors · mean 46
▸ expand
Noise Pattern
25
Frequency Analysis
73
Error Level Analysis
33
Color Distribution
34
Pixel Analysis
57
Compression Quality
48
Edge Consistency
49
Provenance & Metadata3 detectors · mean 54
▸ expand
ICC Profile
62
Metadata
50
C2PA Provenance
50

Image Quality

Dimensions
941 × 1672 px
Aspect
0.563
File size
338.1 KB
Bytes / pixel
0.220

Frequency Analysis

Radial1.000
DCT0.773
Upsampling1.000
Cross-channel0.150
Power-law β
-2.94
Grid energy
0.341

Edge Consistency

CV 0.529
Cell 1: 8.4433Cell 2: 25.1636Cell 3: 26.5894Cell 4: 10.4490Cell 5: 13.6024Cell 6: 47.5626Cell 7: 43.4828Cell 8: 8.1397Cell 9: 19.2080Cell 10: 48.4755Cell 11: 49.7883Cell 12: 15.7636Cell 13: 20.6733Cell 14: 29.0492Cell 15: 45.2516Cell 16: 28.7146

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

Range: 8.1397 – 49.7883

Noise Fingerprint

Variance
189.44
Std deviation
13.76
Mean
-0.0
Spatial corr.
6.630
Mean Δ
2.78
σ
2.78
CV
0.999
Uniformity
0.001

Provenance

Source Dossier

PlatformDirect upload
Author—
Content Typeimage
Analyzed OnMay 13, 2026, 2:03 AM
Analyzed byAnonymous