A practical overview of how neural radiance fields evolved from the original NeRF to compact representations using wavelet transforms and learnable codebooks, and what it means for real-time applications.

How self-supervised sinogram synthesis and geometric refinement can reconstruct high-quality CT volumes from sparse and limited-angle projections — bridging model-based and learning-based approaches.

From real-time 3D capture to reconstruction, transmission, and rendering — a technical deep-dive into the core technologies powering holographic communication systems at Peng Cheng Laboratory.