AI Paper Atlas

Research map / 3D vision / Gaussian splatting

One-Shot Refiner: Boosting Feed-forward Novel View Synthesis via One-Step Diffusion

Yitong Dong, Qi Zhang, Minchao Jiang, Zhiqiang Wu, Qingnan Fan, Ying Feng, Huaqi Zhang, Hujun Bao, Guofeng Zhang

AAAI 2026 · 3D vision › Gaussian splatting · approach: Sparse-view Gaussian splatting for novel views · 6 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 frequency-domain branch (DD-DPM) over CNN-only detail module (ablation)detail perception module · DL3DVFID: 62.60 -> 55.32Improved
add DD-DPM over CNN-only detail module (ablation)detail perception module · DL3DVMUSIQ: 64.95 -> 63.47Hurt
add CNN detail module to base (ablation)base feed-forward model · DL3DVPSNR: 20.08 -> 22.10Improved
add Gaussian features with a 2-layer CNN refiner (ablation)DD-DPM model · DL3DVFID: 55.32 -> 70.51Hurt
add Gaussian features with a 2-layer CNN refiner (ablation)DD-DPM model · DL3DVLPIPS: 0.22 -> 0.27Hurt
add standalone one-step SD refiner (ablation)reconstruction model · DL3DVFID: 70.51 -> 48.58Improved
add standalone one-step SD refiner (ablation)reconstruction model · DL3DVPSNR: 22.22 -> 22.17No clear effect
make SD refiner feature-guided with joint training (ablation)SD refiner · DL3DVFID: 48.58 -> 40.46Improved
make SD refiner feature-guided with joint training (ablation)SD refiner · DL3DVPSNR: 22.17 -> 22.67Improved
use feature-guided SD instead of MARINER enhancementenhancement module · DL3DVFID: 63.04 -> 40.46Improved
use feature-guided SD instead of Difix3D+ SDenhancement module · DL3DVFID: 50.30 -> 40.46Improved
upsample to 1024x1024 with Tile-VAE without retrainingOne-Shot Refiner · DL3DV 1024x1024FID: 108.74 -> 41.74Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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