AI Paper Atlas

Research map / Vision-language and multimodal / Human and embodied modeling

Beyond Semantic Search: Towards Referential Anchoring in Composed Image Retrieval

Yuxin Yang, Yinan Zhou, Yuxin Chen, Ziqi Zhang, Zongyang Ma, Chunfeng Yuan, Bing Li, Jun Gao, Weiming Hu

CVPR 2026 · Vision-language and multimodal › Human and embodied modeling · approach: Composed and text-based image retrieval · 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 CAAM (adaptive attention modulation) to baseline (ablation)AdaFocal · OACIRRAvg recall: 74.91 -> 79.00Improved
use Transformer CRM instead of average pooling (ablation)CAAM · OACIRRAvg recall: 76.39 -> 78.19 (frozen probes)Improved
make probe tokens learnable (Transformer CRM) (ablation)CAAM · OACIRRAvg recall: 78.19 -> 79.00Improved
make probe tokens learnable (average pooling) (ablation)CAAM · OACIRRAvg recall: 76.39 -> 76.31No clear effect
fix a positive modulation scalar beta instead of beta=0 (ablation)AdaFocal · OACIRRRID@1 / R@1Improved
use too large fixed beta (ablation)AdaFocal · OACIRRR@1Hurt
predict beta per query (CAAM) instead of fixed beta (ablation)AdaFocal · OACIRRRID@1 / R@1Improved
train SPRC on OACIRR instead of CIRR (ablation)SPRC (ViT-G) · OACIRRAvg recall: 37.30 -> 74.05Improved
use AdaFocal instead of OACIRR-trained SPRC (ViT-G)AdaFocal · OACIRRRID@1: avg +7.47 RID@1, +4.14 R@1Improved
scale backbone from ViT-L to ViT-GAdaFocal · OACIRRAvg recall: 76.40 -> 79.00Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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