3D scene generation and reconstruction
I combined sparse keypoints, generative reconstruction and optimal transport to reduce the effects of channel errors on 3D scene reconstruction.
I’m Zhe Wang, a PhD candidate at King’s College London. I research wireless communications, generative AI and AI infrastructure.
I combined sparse keypoints, generative reconstruction and optimal transport to reduce the effects of channel errors on 3D scene reconstruction.
I built a federated split-learning testbed with OpenAirInterface and RFsim to measure computation, communication and aggregation costs across virtual devices.
Goal-Oriented Semantic Communications for Avatar-Centric Augmented Reality
I studied skeleton-based avatar updates and joint prioritisation to reduce point-cloud transmission for augmented reality.