ZONN.ai Forensic Report

Case · A3C20288 · 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 ConfidenceLimited · 43/100

Analysed Specimen

Original analysed image
Forensic suspicion heatmap
OriginalHeatmap
POS55/100
No flagged regions

Heads up — 1 thing 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.

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).

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

  • INA v2 (FLUX/MJ)read real · 3/100

    BEiT-Large dual-head classifier trained on FLUX, Midjourney, and real photo corpora.

  • xRayon ConvNeXtV2flagged AI · 97/100

    ConvNeXtV2 detector trained on FLUX, DALL-E 3, SDXL, SD3.5, and Midjourney v6.

Model Agreement

42%

Variance across 6 ML detectors. Higher agreement means the models converged on the same reading; lower agreement means treat the verdict with care.

Evidence — 16 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 Models6 detectors · mean 40
▸ expand
INA v2 (FLUX/MJ)
3
xRayon ConvNeXtV2
97
SigLIP AI Detector
25
CommFor (4803 Generators)
31
Bombek1 SigLIP+DINOv2
32
Manipulation Map (IML-ViT)
50
Pixel & Frequency Forensics7 detectors · mean 52
▸ expand
Error Level Analysis
16
Noise Pattern
84
Pixel Analysis
75
Frequency Analysis
72
Edge Consistency
35
Color Distribution
37
Compression Quality
45
Provenance & Metadata3 detectors · mean 49
▸ expand
ICC Profile
48
Metadata
50
C2PA Provenance
50

Image Quality

Dimensions
3000 × 4000 px
Aspect
0.750
File size
913.9 KB
Bytes / pixel
0.078

Frequency Analysis

Radial1.000
DCT0.809
Upsampling1.000
Cross-channel0.055
Power-law β
-3.10
Grid energy
0.286

Edge Consistency

CV 0.841
Cell 1: 1.0649Cell 2: 2.3071Cell 3: 3.1931Cell 4: 1.1187Cell 5: 1.5121Cell 6: 12.1638Cell 7: 15.8057Cell 8: 1.6006Cell 9: 2.3325Cell 10: 8.2651Cell 11: 10.5032Cell 12: 3.6043Cell 13: 2.9793Cell 14: 13.3694Cell 15: 12.5441Cell 16: 3.0881

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

Range: 1.0649 – 15.8057

Noise Fingerprint

Variance
39.45
Std deviation
6.28
Mean
0.0
Spatial corr.
1.402
Mean Δ
1.11
σ
1.40
CV
1.262
Uniformity
-0.262

Provenance

Source Dossier

PlatformDirect upload
Author—
Content Typeimage
Analyzed OnMay 15, 2026, 11:58 AM
Analyzed byAnonymous