Papers › Data Roaming and Quality Assessment for Composed Image Retrieval
Data Roaming and Quality Assessment for Composed Image Retrieval
Matan Levy, Rami Ben-Ari, Nir Darshan, Dani Lischinski
The task of Composed Image Retrieval (CoIR) involves queries that combine image and text modalities, allowing users to express their intent more effectively. However, current CoIR datasets are orders of magnitude smaller compared to other vision and language (V&L) datasets. Additionally, some of these datasets have noticeable issues, such as queries containing redundant modalities. To address these shortcomings, we introduce the Large Scale Composed Image Retrieval (LaSCo) dataset, a new CoIR dataset which is ten times larger than existing ones. Pre-training on our LaSCo, shows a noteworthy improvement in performance, even in zero-shot. Furthermore, we propose a new approach for analyzing CoIR datasets and methods, which detects modality redundancy or necessity, in queries. We also introduce a new CoIR baseline, the Cross-Attention driven Shift Encoder (CASE). This baseline allows for early fusion of modalities using a cross-attention module and employs an additional auxiliary task during training. Our experiments demonstrate that this new baseline outperforms the current state-of-the-art methods on established benchmarks like FashionIQ and CIRR.
In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.
Code
Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.
Code Syntology ran Syntology
Not run by Syntology. Nothing on this page verifies that the listed code works.
Tasks
Datasets
Introduced by this paper, per the archive.
Results from the paper archive 2025-07-28
| Task | Dataset | Model | Metric | Value | Rank at snapshot | Leaderboard | Report |
|---|---|---|---|---|---|---|---|
| Image Retrieval | CIRR | CASE (Pre-trained on LaSCo.Ca) | (Recall@5+Recall_subset@1)/2 | 78.25 | #7 of 17 | Archive leaderboard | report |
| Image Retrieval | CIRR | CASE (Pre-trained on LaSCo.Ca) | Recall@10 | 88.75 | #7 of 17 | Archive leaderboard | report |
| Image Retrieval | CIRR | CASE | (Recall@5+Recall_subset@1)/2 | 77.5 | #8 of 17 | Archive leaderboard | report |
| Image Retrieval | CIRR | CASE | Recall@10 | 87.25 | #8 of 17 | Archive leaderboard | report |
| Image Retrieval | Fashion IQ | CASE | (Recall@10+Recall@50)/2 | 59.73 | #7 of 22 | Archive leaderboard | report |
| Image Retrieval | Fashion IQ | CASE | Recall@10 | 48.79 | #7 of 22 | Archive leaderboard | report |
| Image Retrieval | LaSCo | CASE | Recall@1 (%) | 7.08 | #1 of 3 | Archive leaderboard | report |
| Image Retrieval | LaSCo | BLIP4CIR | Recall@1 (%) | 4.26 | #2 of 3 | Archive leaderboard | report |
Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.
Methods
Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections