Most of what we want to sense never arrives cleanly. Reconstruction is the inverse problem: inferring the scene from multipath, occlusion and noise — and doing it fast enough to be useful in the field.

The result is a reusable capability across modalities — a building block for the Weibel Group, and, where it earns it, the seed of a company in Weibel Ventures.

Selected directions

Non-line-of-sight reconstruction

Recovering scenes from scattered, indirect returns — seeing around corners.

Multi-sensor tomography

Fusing modalities and viewpoints into one coherent scene (SAR / ISAR).

Learned inverse solvers

AI priors for ill-posed problems, deployable at the edge.

Quantum-accelerated reconstruction

Exploring quantum computing for the large-scale optimisation behind reconstruction.

Interactive — scattering from a target

The inverse problem, made tangible

A plane wave sweeps in from the left and reflects off two unseen targets. What a sensor records is only the scattered field — reconstruction is the work of inferring the shape and position of the targets from those ripples alone. Drag either target to watch the scattered pattern rearrange in real time.

Drag the targets