AI Paper Atlas

Research map / 3D vision / 3D generation

Transformable Bottleneck Networks

Kyle Olszewski, Sergey Tulyakov, Oliver Woodford, Hao Li, Linjie Luo

ICCV 2019 · 3D vision › 3D generation · approach: Diffusion-based novel view synthesis · 82 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

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 segmentation supervision (ablation)TBN novel view synthesis · ShapeNet chair NVS (reduced model)SSIM: 0.920 -> 0.921No clear effect
add segmentation supervision (ablation)TBN novel view synthesis · ShapeNet chair NVS (reduced model)foreground L1: 0.182 -> 0.189No clear effect
synthesize 4 extra views at regular poses and aggregate their bottlenecksTBN occupancy decoder · ShapeNet chair 3D reconstructionIoU: .3042 -> .4507Improved
synthesize 4 extra views at regular poses and aggregate their bottlenecksTBN occupancy decoder · ShapeNet car 3D reconstructionIoU: .4664 -> .7128Improved
synthesize 4 extra views at regular poses and aggregate their bottlenecksTBN occupancy decoder · ShapeNet aeroplane 3D reconstructionIoU: .2699 -> .4661Improved
sample synthesized views at regular instead of random posesTBN 3D reconstruction (+4 views) · ShapeNet chair 3D reconstructionIoU: .3648 -> .4507Improved
sample synthesized views at regular instead of random posesTBN 3D reconstruction (+4 views) · ShapeNet car 3D reconstructionIoU: .5674 -> .7128Improved
use real random-pose views instead of synthesized random-pose viewsTBN 3D reconstruction (+4 views) · ShapeNet chair 3D reconstructionIoU: .3648 -> .3822Improved
add synthesized random-pose viewsTBN 3D reconstruction · ShapeNet chair 3D reconstructionIoU: .3042 -> .3648Improved
increase additional regular synthetic views from 8 to 9TBN 3D reconstruction · ShapeNet chair 3D reconstructionIoU: .4630 -> .4561No clear effect
increase input views at test time from 1 to 4TBN NVS · ShapeNet carL1: .091 -> .059Improved
increase input views at test time from 1 to 4TBN NVS · ShapeNet chairSSIM: .895 -> .939Improved

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