| license | Apache-2.0 |
| name | event-detection-temporal-intelligence-expert |
| description | Expert in temporal event detection, spatio-temporal clustering (ST-DBSCAN), and photo context understanding. Use for detecting photo events, clustering by time/location, shareability prediction, place recognition, event significance scoring, and life event detection. Activate on 'event detection', 'temporal clustering', 'ST-DBSCAN', 'spatio-temporal', 'shareability prediction', 'place recognition', 'life events', 'photo events', 'temporal diversity'. NOT for individual photo aesthetic quality (use photo-composition-critic), color palette analysis (use color-theory-palette-harmony-expert), face recognition implementation (use photo-content-recognition-curation-expert), or basic EXIF timestamp extraction. |
| allowed-tools | Read,Write,Edit,Bash,Grep,Glob,mcp__firecrawl__firecrawl_search,WebFetch |
| category | AI & Machine Learning |
| tags | ["event-detection","temporal","intelligence","pattern-recognition","real-time"] |
| pairs-with | [{"skill":"photo-content-recognition-curation-expert","reason":"Content + temporal understanding"},{"skill":"wedding-immortalist","reason":"Event detection for wedding albums"}] |
Event Detection & Temporal Intelligence Expert
Expert in detecting meaningful events from photo collections using spatio-temporal clustering, significance scoring, and intelligent photo selection.
DECISION POINTS
ST-DBSCAN vs DeepDBSCAN Algorithm Selection
Photo corpus analysis needed?
├─ Have GPS + timestamps?
│ ├─ Same location, different activities detected? ──── DeepDBSCAN
│ │ └─ Cost tolerance: High accuracy > speed ─────── Add CLIP embeddings
│ └─ Simple time/location grouping sufficient? ────── ST-DBSCAN
│
├─ Timestamps only (no GPS)?
│ ├─ Visual similarity important? ───────────────── Temporal + CLIP clustering
│ └─ Pure time-based events? ────────────────────── Temporal binning
│
└─ Need hierarchical events (vacation > daily > moments)?
└─ Multi-level ST-DBSCAN cascade ─────────────────── Expanding ε thresholds
Parameter Selection Matrix
| Event Type | ε_spatial | ε_temporal | min_pts | Use Case |
|---|
| Indoor party | 50m | 4hr | 5 | Home gatherings |
| Wedding | 200m | 8hr | 8 | Venue + reception |
| City tour | 5km | 12hr | 3 | Tourism, exploration |
| Multi-day trip | 50km | 72hr | 10 | Vacation clustering |
| Conference | 1km | 24hr | 6 | Business events |
Event Significance Threshold Decision
Computed significance score?
├─ Score ≥ 0.8? ──────── Life event candidate (birth, wedding, graduation)
├─ Score ≥ 0.6? ──────── Major memorable event
├─ Score ≥ 0.4? ──────── Significant social gathering
├─ Score ≥ 0.2? ──────── Minor event worth keeping
└─ Score < 0.2? ──────── Daily routine, consider filtering
FAILURE MODES
Over-Clustering Syndrome
Symptoms: Every few photos become separate "events"; 50+ micro-events from one vacation
Detection Rule: If >30% of events contain <5 photos AND duration <2 hours
Diagnosis: ε parameters too restrictive, treating natural breaks as separate events
Fix: Increase ε_temporal (2hr → 6hr) or use hierarchical clustering with larger top-level ε
Under-Clustering Collapse
