AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

TIGaussian: Disentangle Gaussians for Spatial-Awared Text-Image-3D Alignment

Jiarun Liu, Qifeng Chen, Yiru Zhao, Minghua Liu, Baorui Ma, Sheng Yang

ICLR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Contrastive vision-language alignment · 0 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 multi-branch 3DGS tokenizer (ablation)UniGS-like baseline · Objaverse classificationtop-1 accuracy: 33.64 -> 35.57Improved
add multi-view image fusion module (ablation)TIGaussian w/ tokenizer + MV · Objaverse image retrievaltop-1 accuracy: 46.19 -> 53.75Improved
add 3D-text projector (ablation)tokenizer-only model · Objaverse text retrievaltext retrieval top-1: 19.15 -> 20.80Improved
add multi-branch tokenizer to full pipeline (ablation)MV+fusion+projector · Objaverse classificationtop-1 accuracy: 37.72 -> 41.76Improved
add multi-view images (no fusion) (ablation)tokenizer model · Objaverse image retrievaltop-1 accuracy: 41.68 -> 46.19Improved
add pretrained Uni3D point-cloud guidance (ablation)3DGS tokenizer · Objaverse-LVIS zero-shottop-1 accuracy: 27.93 -> 41.76Improved
use 12 instead of 6 multi-view images (ablation)TIGaussian · Objaverse-LVIS classificationtop-1 accuracy: 41.76 -> 40.52Hurt
use 1 view vs 6 views (ablation)TIGaussian · Objaverse-LVIS image retrievalimage retrieval top-1: 40.52 -> 52.11Improved
use ground-truth rendered views instead of generated (ablation)TIGaussian · Objaverse-LVIS zero-shottop-1 accuracy: 41.76 -> 41.03No clear effect
use Cap3D captions instead of InternLM (ablation)TIGaussian · Objaverse-LVIS zero-shottop-1 accuracy: 41.76 -> 40.12No clear effect
replace Hunyuan3D-v1-std views with lite (ablation)TIGaussian · Objaverse-LVIS zero-shottop-1 accuracy: 41.76 -> 39.98Hurt
scale training to Objaverse-800kTIGaussian vs UniGS · Objaverse-LVIS zero-shottop-1 accuracy: 48.8 -> 50.1Improved

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