VextoraTech Vision
VextoraTech Vision is a custom-built workforce intelligence product created for SuperAsia Group. It turns existing camera feeds into a touchless attendance and compliance workflow, giving operations teams one place to review attendance, dress-code adherence, and policy events.
Product screens
Operations — live attendance command center
Present, late, unmatched events, and attendance rate update as people move through the gates. Flow charts and department bars show where the day is healthy or slipping. Attention items flag offline cameras, unmatched recognitions, and pending corrections. This is the screen operations teams use to run the floor, not review footage later.
Live monitor — cameras and recognition stream
A multi-camera wall shows online, degraded, and offline feeds across gates and production areas. The event stream separates matches, low-confidence hits, unmatched faces, and spoof attempts. Confidence scores and timestamps keep every recognition reviewable. Proves the product watches existing cameras in real time, not just a daily spreadsheet.
People — workforce identities and enrollment
Employee records span departments, sites, status, and face-enrollment state in one directory. Bilingual names and site filters match a multi-plant SuperAsia workforce. Import and add-employee actions keep enrollment operational rather than a one-off setup. This is the controlled identity layer behind recognition accuracy.
Daily sheet — punches, lateness, and exceptions
Each employee-day shows first in, last out, worked hours, overtime, and late or early-leave flags. Source marked as face makes the attendance trail auditable. Manual entry, export, and finalise-day actions support HR close-out. Turns camera events into the attendance sheet teams already need for payroll and corrections.
Shifts — grace, thresholds, and night rules
Shift definitions encode hours, grace windows, half-day and full-day thresholds, breaks, and overtime triggers. Night and graveyard shifts that cross midnight are modeled explicitly. Weekday patterns keep rosters aligned with plant reality. Without this layer, recognition events cannot become fair attendance outcomes.
The challenge
- Manual attendance processes create queues, corrections, and opportunities for proxy check-ins.
- A production environment needs reliable face detection across changing light, camera angles, movement, and crowded entry points.
- Dress-code and workplace rules are difficult to monitor consistently across shifts and locations.
- Operations teams need searchable evidence and clear reports rather than another stream of unstructured camera footage.
Our solution
- Built a real-time detection and recognition pipeline that identifies enrolled employees from live camera feeds.
- Connected recognition events to attendance rules for automated check-ins, check-outs, and exception review.
- Added configurable visual checks for dress-code and compliance policies without interrupting normal movement.
- Designed an operations dashboard for enrollment, daily attendance, flagged events, evidence review, and reporting.
- Separated detection, identity matching, and business rules so each layer can be monitored and improved independently.
Key features
Touchless Attendance
Recognizes enrolled employees at entry points and records attendance without cards, fingerprints, or manual registers.
Dress-Code Monitoring
Checks configurable uniform and workplace attire requirements, then routes exceptions for authorized review.
Compliance Events
Creates time-stamped policy events with camera context so teams can review exceptions instead of watching full recordings.
Employee Enrollment
Provides a controlled workflow for adding employee identities, managing reference images, and updating workforce records.
Attendance Reporting
Surfaces daily attendance, late arrivals, missed events, and exceptions in searchable operational reports.