.png)
Attendance Tracking
Real-time attendance logging with face recognition in VisionLog - AI Attendance
The VisionLog - AI Attendance module provides real-time attendance tracking using face recognition with automatic logging to PostgreSQL database.
Features
- Real-time Recognition - Live camera feed with instant face detection
- Automatic Logging - Records attendance with timestamp and confidence to database
- Duplicate Prevention - Prevents multiple entries for the same person
- Database Storage - All records stored in PostgreSQL for reliability and querying
Live Attendance Mode
Display Information
The live view shows:
- Green box: Person just logged
- Orange box: Person already logged
- Red box: Unknown person
- Status bar: Present count and total logs
Keyboard Controls
| Key | Action |
|---|---|
ESC / Q | Quit and show summary |
S | Show current attendance |
R | Reload face database |
Duplicate Prevention
The system prevents duplicate attendance entries for the same person within a session.
Visual feedback:
LOGGED- Just logged successfullyAlready logged- Already recorded in this session
Attendance Record
Each attendance record stored in PostgreSQL contains:
| Field | Description |
|---|---|
| ID | Unique record identifier |
| Person ID | Reference to enrolled person |
| Timestamp | Date and time of attendance |
| Camera ID | Source camera identifier |
| Confidence | Recognition confidence (0-1) |
| Snapshot Path | Path to face snapshot (optional) |
Viewing Attendance
Query attendance records from the database:
- Filter by date range
- Filter by person
- Filter by camera
- Sort by timestamp
Exporting Attendance
Export attendance records for reporting:
- Export to CSV format
- Export specific date range
- Generate summary reports
Session Summary
When you end an attendance session, the system displays:
- Session date
- Total persons present
- Total log entries
- List of all attendees
Attendance Tracker Functions
| Function | Description |
|---|---|
| Check Log Status | Verify if person can be logged |
| Log Attendance | Record attendance entry to database |
| Get Today's Records | Query today's entries from database |
| Get Present Names | List of unique names present today |
| Get Summary | Statistics for the session |
Video Processing Integration
Process recorded video files to extract attendance:
- First-seen time as check-in
- Last-seen time as check-out
- Total appearance count per person
Database Configuration
| Setting | Description |
|---|---|
| Host | PostgreSQL server address |
| Port | Database port (default: 5432) |
| Database | Database name |
| User | Database username |
| Password | Database password |
Status Indicators
Face Box Colors
| Color | Meaning |
|---|---|
| Green | Just logged |
| Orange | Already logged |
| Red | Unknown person |
On-Screen Stats
Displays:
- Present count for today
- Total log entries
- Number of faces currently in frame
Troubleshooting
Person not being logged
- Check if already logged (orange box)
- Verify person is enrolled
- Check recognition confidence
Duplicate entries
- Ensure camera angle is consistent
- Check for multiple face detections
Missing entries
- Lower similarity threshold if needed
- Re-enroll person with better images
- Check lighting conditions
Database connection issues
- Verify PostgreSQL is running
- Check connection credentials
- Ensure network connectivity
Best Practices
Camera Setup
- Position at eye level
- Ensure good lighting
- Avoid backlighting
- Single entry/exit point works best
Data Management
- Regular database backups
- Archive old attendance records
- Review unknown faces for enrollment