Neuromorphic vision sensors entered public-space deployments with a structural advantage over conventional cameras: their output — sparse, asynchronous event streams rather than full image frames — appeared inherently resistant to identity reconstruction. That appearance was never formally tested against the adversarial case. Müller, Kocsis, and Stache tested it . The result is that existing obfuscation methods — masking and scrambling — corrupt the spatio-temporal structure badly enough to degrade downstream perception tasks, while leaving identity exposure unaddressed at the generative reconstruction layer. The paper's contribution is a framework that resolves…