6G is expected to deliver not only multi-gigabit speeds, but it will also expose an image of the world via integrated sensing and communications, which can convert ordinary communication signals into perception. Such conversions must hold up in noise, latency must stay low for real-time digital twins, and information must be used from different viewpoints to reliably construct camera-like images.
We’re introducing a simple two-segment optimizer that enables the first multi-bounce ISAC imaging at 6.75 GHz that can transform standard channel state information (CSI) into point-cloud reconstructions of real objects, such as tree, cubes, and vehicles, which is obtained by fusing multiple TX/RX views and validated on NYURay ray tracing.

In this project, we build a physics-guided ISAC imaging pipeline that turns standardized CSI into dense 3D RF maps via equivalent reflection points and multi-vantage fusion, where we have:

• NYURay-to-CSI pipeline at 6.75 GHz, enabling the first multi-bounce ISAC imaging demonstration at this band.
• Multi-vantage fusion with geometric filtering from different Tx/Rx viewpoints.
• Interface support in MATLAB and ability to import 3D models from softwares like Blender.