Abstract
Public spaces, including shopping centres, restaurants, banks and offices, must grapple with how to offer easy access to the public while ensuring occupant safety and hygiene. We present SafeSpace, a modular smart-environment platform that gives administrators control over indoor density using appointment-based access, enforces physical distancing through proximity alerts in real-time, and verifies PPE adherence with on-device deep-learning mask detection. The system was developed using YOLOv3 and MobileNetV2 on Intel Core i5 CPU and Google Colab, and obtained 98% precision, recall, and F1-score on a real-world dataset. Real-time proximity alerts were implemented by employing YOLOv3 and ROI-based centroid distance estimation with an overall latency of 125–158 ms per frame. In addition, the simulated entry flow comparisons between walk-ins and a predetermined schedule indicated that peak occupancy was reduced by 38.24% during a 15-min interval. SafeSpace’s modular design and consolidated analytics dashboard also make it leveraged for other public-space scenarios and the next generation of post-pandemic facility management and day-to-day operational safety.
| Original language | English |
|---|---|
| Pages (from-to) | 316-331 |
| Number of pages | 16 |
| Journal | International Journal of Advances in Soft Computing and its Applications |
| Volume | 17 |
| Issue number | 3 |
| DOIs | |
| Publication status | Published - 2025 |
All Science Journal Classification (ASJC) codes
- Computer Science Applications
Fingerprint
Dive into the research topics of 'Open for business, closed for crowding: SafeSpace a Smart Solution'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver