Video Anomaly Detection — Interactive Stakeholder Guide

An interactive decision guide for evaluating and deploying CCTV anomaly detection models. Walk through a structured evaluation from model comparison to deployment recommendation.

Interactive Guide
VAD Model Evaluation — Stakeholder Decision Guide
Stage 1 of 5Model Comparison

Model Comparison

You have 5 model iterations from an ablation study. Before investigating details, compare them side-by-side to identify which candidate is worth pursuing.

Why This Matters

This is where you decide if there's a candidate worth investigating. If no model improves meaningfully over Baseline, stop here and re-evaluate the architectural approach.

Diagnostic Task

Toggle models on and off in the comparison strip. Compare F1, FAR, and Confirm Rate to identify the strongest candidate.

What to Look For
  • F1 score improvements — does the candidate improve by more than 5% over Baseline?
  • FAR/100h — a high F1 with unacceptable FAR means operators will be overwhelmed
  • Confirm Rate — if operators only confirm 60% of alerts, the model may be generating nuisance alerts
Checkpoints
Select Candidate Model (A4) in comparison list
Keep Baseline selected to view delta comparison
You should leave this stage knowing which model (if any) you want to evaluate for deployment.
navigate
Active Workspace: Ablation Study Comparison
Models:
ModelPR-AUCF1tIoUPrecisionRecallFAR/100hConfirm RateFrag. IndexLatency (s)Status
Baseline (Frame-level)
0.720
0.680
0.420
70.0%
65.0%
14.2
61%
3.40
2.8s
completed
A1: Memory-guided
+6.00.780
+6.00.740
+13.00.550
+6.076.0%
+7.072.0%
-1.412.8
+4.065%
-0.62.80
-0.42.4s
completed
A2: Temporal Smoothing
+10.00.820
+11.00.790
+22.00.640
+12.082.0%
+11.076.0%
-3.111.1
+10.071%
-1.71.70
-0.72.1s
completed
A3: Event Persistence
+13.00.850
+15.00.830
+29.00.710
+16.086.0%
+15.080.0%
-4.89.4
+15.076%
-2.31.10
-1.01.8s
completed
A4: Full Stack (Candidate)candidate
+17.00.890
+19.00.870
+36.00.780
+21.091.0%
+20.085.0%
-7.46.8
+23.084%
-2.31.05
-1.41.4s
completed