Detection Metrics
Transparent benchmarks so you know exactly how Cacuda performs. Tested on a mixed dataset of AI-generated and human-created content.
87%
Accuracy
72.7%
Precision
100%
Recall
0.842
F1 Score
Ensemble vs. Individual Models
The Cacuda Score outperforms every single model in isolation.
| Detector | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|
| Cacuda Score | 87.0% | 72.7% | 100% | 0.842 |
| Ensemble (SigLIP2 + DINOv2) | 73.9% | 57.1% | 100% | 0.727 |
| NPR (Frequency Domain) | 65.2% | 50.0% | 87.5% | 0.636 |
| SPAI (Spectral Patch) | 78.3% | 100% | 37.5% | 0.545 |
Model Architecture
Five complementary detection techniques working together.
Ensemble
Weight: 0.45SigLIP2 + DINOv2 with LoRA. Dual-encoder combining semantic and self-supervised features.
45%
NPR
Weight: 0.30ResNet-50 analysing Neighbouring Pixel Relationships in the frequency domain.
30%
SPAI
Weight: 0.25Vision Transformer with Masked Feature Modelling for spectral patch-based learning.
25%
DIRE
Bonus: +0.05ResNet-50 measuring Diffusion Reconstruction Error. Secondary confirmation signal.
+5%
PatchCraft
Bonus: +0.05SRM-CNN for texture patch analysis using the Smash & Reconstruct technique.
+5%
Score Interpretation
How to read the Cacuda Score.
> 0.7
Likely AI-Generated
0.5 – 0.7
Possibly AI-Generated
0.3 – 0.5
Probably Real
< 0.3
Likely Real