SJTU's GeoVerse Synthesizes World-Consistent Novel Views in Geometric Latent Space
SJTU · hf · 2026-09-29
SJTU proposes GeoVerse, which performs generation inside the geometric latent space of a pretrained 3D foundation model while injecting appearance priors from Wan2.2 VACE via a ControlNet-style adapter; a global spatial memory anchors cross-view coherence. It improves visual quality and geometric consistency, with 2.23 dB higher PSNR on DL3DV and 32.4% lower ATE on Mip-NeRF360 versus GLD.
More from Research
- NVIDIA's LSPD brings RL tricks to policy distillation, cutting rollouts by 75% — nvidia · 2026-09-29
- Swapping matmul for associative-algebra layers boosts 110M LM throughput 7.8% — Ilya Koziev · 2026-09-29
- SMAT: merge-aware training lifts merged model scores up to 2.16 with <2% overhead — PolyUHK · 2026-09-29
- KernelZero-7B co-evolution beats Claude 4.5 Sonnet on CUDA kernel generation — Changxin Ke · 2026-09-29
- Frozen-base 34M logit correction module fixes 53.3% of Gemma errors losslessly — eulogik · 2026-09-29
- 6.5M-param NanoForecast beats 200M TimesFM on ETT after pipeline fixes cut MASE 43.8% — eulogik · 2026-09-29