PhotoBench is a new benchmark built from authentic personal photo albums that moves beyond simple visual matching to focus on personalized, intent-driven retrieval. It incorporates a multi-source profiling framework that combines visual semantics, spatial‑temporal metadata, social identity, and temporal events to generate complex queries reflecting users’ life trajectories. Evaluation on PhotoBench reveals two key limitations: a modality gap where unified embedding models fail on non‑visual constraints, and a source fusion paradox where agentic systems struggle with tool orchestration.
By Tianyi Xu, Rong Shan, Junjie Wu, Jiadeng Huang, Teng Wang, Jiachen Zhu, Wenteng Chen, Minxin Tu, Quantao Dou, Zhaoxiang Wang, Changwang Zhang, Weinan Zhang, Jun Wang, Jianghao Lin
The paper introduces PRISm, an image-to-image retrieval framework that prioritizes important elements in scene graphs for semantic image search. PRISm uses an Importance Prediction Module, trained with caption-derived relevance, to prune irrelevant objects and relations, and an Edge-Aware Contextual GNN to fuse the refined graph with global visual features into a single embedding. Experiments on PSG and Flickr show that PRISm outperforms strong baselines in MAP@1 and top‑1 metrics while adding only 33 ms per query beyond scene‑graph generation.
By Dimitrios Georgoulopoulos, Nikolaos Chaidos, Angeliki Dimitriou, Giorgos Stamou
Current instruction-based image retrieval systems are powerful but limited to single-turn interactions, failing to capture the iterative nature of complex, real-world visual searches. To overcome this limitation, we introduce Contextual Composed Image Retrieval (CoCo-IR), a novel task that enables users to progressively refine search results through interactions.
CORE improves compositional reasoning in multimodal language models by distilling a cross‑attentive reranker’s fine‑grained judgments into the embedding model. It generates candidate lists across five compositional matching levels and trains with a Rank‑KL objective to replicate the reranker’s ranking. Experiments on COLA, SUGARCREPE++, and NEGBENCH show CORE‑RERANKER‑8B outperforms Jina‑Reranker by 10.7 points, while CORE‑EMBED‑8B achieves the best overall average among evaluated embeddings, with gains also transferring to the MCMR benchmark without harming COCO or Flickr30K retrieval.
By Tingyu Song, Mingxin Li, Yanzhao Zhang, Dingkun Long, Chu Liu, Pengjun Xie, Yilun Zhao, Shu Wu
MulVec is a training‑free zero‑shot composed image retrieval method that matches a target image to a gallery using a reference image and a text edit. It introduces a role‑aware query system that separates the target description into four retrieval roles—Global, Desired, Preserve, and Forbidden—each mapped to specific probe vectors. By combining global and local visual representations, MulVec achieves significant performance gains on CIRCO, CIRR, and FashionIQ datasets, improving CIRCO mAP@5 by up to 23.0% over prior methods.
By Zihao Zhang, Dayan Wu, Xinze Liu, Hengjie Zhu, Yiliang Zhu, Ding Wang, Peng Fu, Zheng Lin, Weiping Wang
arXiv:2609.10008v1 Announce Type: new
Abstract: Composed video retrieval (CoVR) searches a gallery for the target video that realizes a natural-language modification of a source clip. However, at gal...
By Dmitry Demidov, Muhammad Zaigham Zaheer, Omkar Thawakar, Abdelrahman Mohamed Shaker, Rao Anwer
arXiv:2606. 31222v1 Announce Type: new Abstract: Composed image retrieval requires identifying a target image from a gallery by integrating a reference image with a textual modification instruction.
By Gunho Jung, Jeong-Woo Park, Seon Bin Kim, Seong-Whan Lee
arXiv:2510. 01483v3 Announce Type: replace-cross Abstract: Vision-language models (VLMs) demonstrate strong image-level scene understanding, but reasoning over long egocentric video remains costly: because VLMs maintain no persistent memory or explicit spatial representation, all sampled frames must be re-processed for every new query.
By Mohamad Al Mdfaa, Svetlana Lukina, Timur Akhtyamov, Arthur Nigmatzyanov, Dmitrii Nalberskii, Sergey Zagoruyko, Gonzalo Ferrer
The paper introduces PRAISE, a test‑time adaptation framework for vision‑language model based person re‑identification. PRAISE constructs an Attribute‑Conditioned Subspace (ACS) from a bank of attribute prompts, reshaping query‑gallery affinities without requiring identity or attribute labels. Experiments on image‑to‑image and text‑to‑image ReID benchmarks show that PRAISE consistently improves retrieval performance over existing backbones and TTA baselines.
By Rajarshi Bhattacharya, Aryan Shukla, Shakeeb Murtaza, Christian Desrosiers, Jose Dolz, Maguelonne Heritier, Eric Granger
The paper proposes augmenting the global [CLS] token and register tokens of vision transformers with spatial region tokens to better capture fine‑grained image details. By matching each semantic token with a corresponding image patch, the method automatically generates localized ROI tokens without external annotations. These tokens are then used in a multi‑vector retrieval framework inspired by ColBERT, improving fine‑grained image retrieval while keeping storage costs manageable.
By Ian de Holanda Cavalcanti Bezerra, Vivek Trivedy, Lucas Pascotti Valem, Longin Jan Latecki
arXiv:2607. 00374v1 Announce Type: cross Abstract: Composed Image Retrieval (CIR) retrieves a target image from a reference image and a textual modification.
By Jingjing Zhang, Lei Zhang, Zheren Fu, Zhendong Mao
arXiv:2604. 09508v2 Announce Type: replace-cross Abstract: Visual Retrieval-Augmented Generation (VRAG) empowers Vision-Language Models to retrieve and reason over visually rich documents.
By Yucheng Shen, Jiulong Wu, Jizhou Huang, Dawei Yin, Lingyong Yan, Min Cao