Deeptrack Logo
Deeptrack
Gotham back office
Client view
OverviewScan logAPI usageCredits & billingClient accounts
Internal
FP / FN queueModel feedbackScan forensicsDataset pipeline
System
API keysWebhooksAudit logAlertsSettings
AD
Admin
Admin
Model feedback
Accuracy trends, FP/FN rates, retraining queue, dataset pipeline
Model accuracy
94.1%
−2.4pts in 12 weeks
FP rate
2.8%
+0.7pts — rising
FN rate
3.1%
+1.7pts — rising
Labels queued
14
Next export: 02:00 EAT
Avg confidence
89.1%
−3.2pts audio drift
Model accuracy & confidence trend
⚠ Drift detected
FP / FN rates over time
Accuracy by media type
Media typeAccuracyFP rateFN rateAuto-escalate thresholdTotal samples
Video
94.5%
2.1%3.4%Escalate below 75%1,840
Audio
93.3%
3.8%2.9%Escalate below 80%1,120
Image
96.8%
1.4%1.8%Escalate below 70%980
Document
97.8%
1%1.2%Escalate below 65%620
Label pipeline
Next export: 02:00 EAT
Scan IDLabelReviewerCertaintyStatus
sc_7c4d9e1bauthenticTauil M.certainqueued
sc_4d7a0c3eauthenticBrian K.likelyaugmented
sc_2c3d4e5fdeepfakeTauil M.certainaugmented
sc_5e6f7a8bdeepfakeBrian K.likelytrained
Retrain history
Run IDDateSamplesOutcomeDuration
rt_0042026-03-284,820+1.2% accuracy6h 14m
rt_0032026-02-143,640+0.8% accuracy5h 02m
rt_0022026-01-102,100+2.1% accuracy4h 38m
rt_0012025-12-011,200Baseline3h 55m
Last retrain: 2026-03-28 · Next scheduled: 2026-04-28 (pending 14 queued labels)