VIRD: View-Invariant Representation through Dual-Axis Transformation for Cross-View Pose Estimation | Signal Canvas | ScienceToStartup