AI Paper Atlas

Research map / Point clouds / Shape and surface completion

4Diff: 3D-Aware Diffusion Model for Third-to-First Viewpoint Translation

Feng Cheng, Mi Luo, Huiyu Wang, Alex Dimakis, Lorenzo Torresani, Gedas Bertasius, Kristen Grauman

ECCV 2024 · Point clouds › Shape and surface completion · approach: Point cloud diffusion generation · 12 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add 3D-aware rotary cross-attention (ablation)4Diff · Ego-Exo4D-VT testLPIPS: 0.373 -> 0.349Improved
add egocentric point cloud rasterization prior (ablation)4Diff · Ego-Exo4D-VT testLPIPS: 0.412 -> 0.373Improved
use pretrained MiDaS depth estimator (ablation)4Diff · Ego-Exo4D-VT testLPIPS: 0.392 -> 0.349Improved
use DINOv2 feature encoder instead of CLIP (ablation)4Diff · Ego-Exo4D-VT testLPIPS: 0.366 -> 0.349Improved
replace geometry-free diffusion (3DiM) with 4Diff geometry priors4Diff · Ego-Exo4D-VT testLPIPS: 0.385 -> 0.349Improved
evaluate on unseen scenes instead of seen scenes4Diff · Ego-Exo4D-VT unseen scenesPSNR / SSIM: PSNR 14.45 vs 3DiM 14.90; SSIM 0.508 vs 0.542Mixed
use NeRF-based GNT for geometry instead of diffusion modelsGNT baseline · Ego-Exo4D-VT testLPIPS: 0.482 vs 0.349 for 4DiffHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy