AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

3D CoCa v2: Contrastive Learners with Test-Time Search for Generalizable Spatial Intelligence

H. W. Tang, Ting Huang, Zeyu Zhang

IJCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Spatial reasoning in VLMs · 5 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 test-time search (best-of-8, LLM judge) to 3D CoCa3D CoCa · ScanRefer (w/o 2D, IoU 0.25)CIDEr: 85.42 -> 86.95Improved
add test-time search to 3D CoCa3D CoCa · ScanRefer (w/o 2D, IoU 0.50)CIDEr: 77.13 -> 78.63Improved
add test-time search to 3D CoCa3D CoCa · Nr3D (IoU 0.5)CIDEr: 52.84 -> 54.45Improved
add test-time search to 3D CoCa3D CoCa · TOD3Cap (zero-shot OOD, IoU 0.5)CIDEr: 47.2 -> 50.8Improved
add test-time search to 3D CoCa3D CoCa · TOD3Cap (zero-shot OOD, IoU 0.25)CIDEr: 55.8 -> 59.6Improved
add test-time search to 3D CoCa3D CoCa · ScanRefer (w/ 2D, IoU 0.50)CIDEr: 74.52 -> 75.60Improved
add contrastive loss (lambda 0 -> 1.0) (ablation)3D CoCa backbone · ScanRefer (IoU 0.25, no TTS)CIDEr: 74.12 -> 85.42Improved
raise contrastive weight from 1.0 to 2.0 (ablation)3D CoCa backbone · ScanRefer (IoU 0.25, no TTS)CIDEr: 85.42 -> 76.89Hurt
use CoCa-style multimodal decoder instead of GPT-2 captioner (ablation)3D CoCa backbone · ScanRefer (IoU 0.25, no TTS)CIDEr: 76.20 -> 85.42Improved
use proposed point-tokenizer + frozen CLIP encoder instead of PointNet++ (ablation)3D CoCa backbone · ScanRefer (IoU 0.25, no TTS)CIDEr: 72.48 -> 85.42Improved
use GPT-5 judge instead of Qwen3-VL Pro (ablation)3D CoCa v2 TTS · ScanRefer (IoU 0.5)CIDEr: 78.00 -> 78.63Improved
use GPT-5 judge instead of Qwen3-VL Pro (ablation)3D CoCa v2 TTS · TOD3Cap (zero-shot OOD, IoU 0.5)CIDEr: 49.20 -> 50.80Improved

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