Double-Helix Vision (DH-V2): A Geometry-Based Visual Sampler for Bandwidth-Constrained Perception

Double-Helix Vision (DH-V2) introduces a geometry-based visual sampling technique that compresses 2D images into 1D signals using spiral trajectories inspired by the golden ratio, achieving a 1,433x compression ratio at 4K resolution with minimal geometric distortion. The system operates at 0.52 ms for perception tasks on CPU-only hardware and shows a 6.03% accuracy improvement over uniform sampling on CIFAR-10 with a sampling budget of 128 points per helix. This approach is significant for practitioners as it offers a lightweight, efficient method for spatial perception in bandwidth-constrained environments, with an accessible API for integration into robotics applications.