🧠 Fractus Scientific Demo

Per-image live benchmark β€” Mapillary 5k (500-frame loop, 2 s / frame)

Mapillary sample
πŸ–ΌοΈ β€”
πŸ“ GT: β€”
πŸͺͺ Plonk err: β€”
🧬 Fractus err: β€”
⏱️ Frame: 0 / 500

⚑ Fractus Engine β€” Real-Time Stability Monitor

Average Error (km)

Plonk vs Fractus (per image)
Plonk: β€”
Fractus: β€”
Gain: β€”

Entropy & Coherence

Derived from stability window
Entropy: β€”
Coherence: β€”
πŸŒ€ Entropy: β€”
πŸ”— Coherence: β€”
⚑ Efficiency: β€”
πŸ“ˆ Stability: β€”

Fractal Stability β€” Std Dev (km)

Rolling window (60)
Plonk Οƒ: β€”
Fractus Οƒ: β€”

Processing Efficiency

Relative accuracy ratio
Efficiency: β€”
Status: Live (2 s)

Benchmark Logs β€” Anomaly Detection on 50 000 Entries

Detection Rate (%)

False Positive Rate (%)

Mean Detection Latency (ms)

Accuracy / Throughput