The July 2026 edition of the Proceedings of the Nigerian Academy of Science is now online!
Explore cutting-edge research from leading scientists across diverse fields and access the full issue here: https://t.co/tYEN8GLFog
#PNgAS#NAS#OpenAccess#Research#Science#Nigeria
Most Re-ID systems assume GPU hardware. That cuts out a huge portion of the world.
Our new paper fixes that.
"Adaptive Frame Sampling for CPU-Based Multi-Camera Person Re-ID" — published today in DRJEIT (Vol. 14, 2026)
delivers real-time multi-camera person tracking on standard CPU hardware, no GPU needed.
The numbers:
- 89.3% fewer inference frames
- 71.1% faster processing
- Up to 56.5 FPS on CPU
- 100% ID consistency, zero switches
Months of work. Fully deployable. Built for the real world