Hello NVIDIA Community Team,
I am the lead researcher at SyntherIQ Labs, an independent research studio focused on edge AI and digital wellness. I am currently developing IrisGuard, a real-time biometric system designed to detect digital fatigue and eye strain via eye-dilation analysis.
The Architecture: My pipeline treats incoming RTSP streams from IP cameras not as video, but as a Raw Data Pipeline. Every frame is decomposed immediately into high-dimensional numerical tensors. This “zero-copy” approach allows me to perform deep mathematical analysis on eye landmarks without ever storing visual images, ensuring absolute user privacy.
The Hardware Bottleneck: I am currently prototyping this architecture on an Enterprise CPU-based Home Lab. While the mathematical logic is solid, the computational overhead of decomposing 2 million+ pixels per frame into raw math is maxing out my CPU. This prevents the sub-30ms latency required for accurate, real-time biometric feedback.
The Goal: I am seeking a Jetson Orin Nano sponsorship to transition SyntherIQ Labs from a CPU-constrained prototype to a dedicated hardware-accelerated edge solution.
Thank you for your consideration and for supporting independent innovation!
Best regards,
OseeRelafordIII/SyntherIQ Labs