{"about":{"non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","site":"https://codewithpapers.app","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page","syntology":{"site":"https://syntology.ai","developers":"https://syntology.ai/developers","mcp":{"server":"https://syntology.ai/mcp","transport":"streamable-http","server_card":"https://syntology.ai/.well-known/mcp/server-card.json","auth":{"type":"trial token, no account","trial_token":"https://syntology.ai/api/oauth/trial/token","method":"POST","docs":"https://syntology.ai/developers"}},"have":"https://syntology.ai/api/graph/have?x=<method, arXiv id or title> (free, answers coverage only)","paper_base":"https://syntology.ai/paper/","atlas_base":"https://app.syntology.ai/?focus="},"machine_readable":[{"url":"https://codewithpapers.app/llms.txt","what":"the machine catalog: every machine-readable file, counted"},{"url":"https://codewithpapers.app/index/manifest.json","what":"paper-to-code index by arXiv id, with Syntology's counts"},{"url":"https://codewithpapers.app/search/manifest.json","what":"site search index (titles, authors) and its files"},{"url":"https://codewithpapers.app/download","what":"bulk files: Syntology's layer, described there"},{"url":"https://codewithpapers.app/build_manifest.json","what":"the build record: inputs, counts, exclusions, probes"}]},"url":"/paper/zero-shot-composed-image-retrieval-with","title":"Zero-Shot Composed Image Retrieval with Textual Inversion","arxiv_id":"2303.15247","date":"2023-03-27","proceeding":"ICCV 2023 1","authors":["Alberto Baldrati","Lorenzo Agnolucci","Marco Bertini","Alberto del Bimbo"],"abstract":"Composed Image Retrieval (CIR) aims to retrieve a target image based on a query composed of a reference image and a relative caption that describes the difference between the two images. The high effort and cost required for labeling datasets for CIR hamper the widespread usage of existing methods, as they rely on supervised learning. In this work, we propose a new task, Zero-Shot CIR (ZS-CIR), that aims to address CIR without requiring a labeled training dataset. Our approach, named zero-Shot composEd imAge Retrieval with textuaL invErsion (SEARLE), maps the visual features of the reference image into a pseudo-word token in CLIP token embedding space and integrates it with the relative caption. To support research on ZS-CIR, we introduce an open-domain benchmarking dataset named Composed Image Retrieval on Common Objects in context (CIRCO), which is the first dataset for CIR containing multiple ground truths for each query. The experiments show that SEARLE exhibits better performance than the baselines on the two main datasets for CIR tasks, FashionIQ and CIRR, and on the proposed CIRCO. The dataset, the code and the model are publicly available at https://github.com/miccunifi/SEARLE.","url_abs":"https://arxiv.org/abs/2303.15247v2","url_pdf":"https://arxiv.org/pdf/2303.15247v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"zero-shot-composed-image-retrieval-with","repo_url":"https://github.com/miccunifi/circo","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"zero-shot-composed-image-retrieval-with","repo_url":"https://github.com/miccunifi/searle","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"composed-image-retrieval","task_name":"Composed Image Retrieval (CoIR)"},{"task_slug":"image-retrieval","task_name":"Image Retrieval"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":"zero-shot-composed-image-retrieval-zs-cir","task_name":"Zero-Shot Composed Image Retrieval (ZS-CIR)"}],"methods":[{"method_slug":"clip","method_name":"CLIP"}],"datasets_introduced":[{"slug":"circo","name":"CIRCO","full_name":"Composed Image Retrieval on Common Objects in context"}],"methods_introduced":[],"results":[{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"CIRCO","model":"SEARLE-XL (CLIP L/14)","rank_in_archive_order":35,"of":43,"metrics":{"mAP@10":"12.73"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"CIRCO","model":"SEARLE (CLIP B/32)","rank_in_archive_order":40,"of":43,"metrics":{"mAP@10":"9.94"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-1","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"CIRR","model":"SEARLE","rank_in_archive_order":43,"of":47,"metrics":{"R@5":"53.42"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-1","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"CIRR","model":"SEARLE-XL","rank_in_archive_order":45,"of":47,"metrics":{"R@5":"52.48"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-4","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"COCO (Common Objects in Context)","model":"SEARLE-XL-OTI (CLIP L/14)","rank_in_archive_order":2,"of":10,"metrics":{"Actions Recall@5":"31.43"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-4","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"COCO (Common Objects in Context)","model":"SEARLE-XL (CLIP L/14)","rank_in_archive_order":4,"of":10,"metrics":{"Actions Recall@5":"29.02"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-4","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"COCO (Common Objects in Context)","model":"SEARLE-OTI (CLIP B/32)","rank_in_archive_order":8,"of":10,"metrics":{"Actions Recall@5":"26.00"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-4","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"COCO (Common Objects in Context)","model":"SEARLE (CLIP B/32)","rank_in_archive_order":10,"of":10,"metrics":{"Actions Recall@5":"24.58"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-2","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"Fashion IQ","model":"SEARLE-XL-OTI (CLIP L/14)","rank_in_archive_order":31,"of":41,"metrics":{"(Recall@10+Recall@50)/2":"37.76"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-2","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"Fashion IQ","model":"SEARLE-XL (CLIP L/14)","rank_in_archive_order":34,"of":41,"metrics":{"(Recall@10+Recall@50)/2":"35.90"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-2","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"Fashion IQ","model":"SEARLE (CLIP B/32)","rank_in_archive_order":38,"of":41,"metrics":{"(Recall@10+Recall@50)/2":"32.71"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-2","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"Fashion IQ","model":"SEARLE-OTI (CLIP B/32)","rank_in_archive_order":39,"of":41,"metrics":{"(Recall@10+Recall@50)/2":"32.39"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-3","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"FashionIQ","model":"SEARLE-XL-OTI","rank_in_archive_order":1,"of":1,"metrics":{"R@10":"27.61"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-11","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"GeneCIS","model":"SEARLE (CLIP B/32)","rank_in_archive_order":4,"of":11,"metrics":{" A-R@1":"14.4"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-11","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"GeneCIS","model":"SEARLE (CLIP L/14)","rank_in_archive_order":5,"of":11,"metrics":{" A-R@1":"14.4"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-5","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"ImageNet","model":"SEARLE-XL (CLIP L/14)","rank_in_archive_order":4,"of":11,"metrics":{"Average Recall":"21.54"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-5","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"ImageNet","model":"SEARLE-XL-OTI (CLIP B/32)","rank_in_archive_order":5,"of":11,"metrics":{"Average Recall":"20.42"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-5","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"ImageNet","model":"SEARLE-OTI (CLIP B/32)","rank_in_archive_order":10,"of":11,"metrics":{"Average Recall":"12.77"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-5","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"ImageNet","model":"SEARLE (CLIP B/32)","rank_in_archive_order":11,"of":11,"metrics":{"Average Recall":"11.94"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-6","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"ImageNet-R","model":"SEARLE-XL (CLIP L/14)","rank_in_archive_order":7,"of":20,"metrics":{"(Recall@10+Recall@50)/2":"21.54"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-6","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"ImageNet-R","model":"SEARLE-XL-OTI (CLIP B/32)","rank_in_archive_order":8,"of":20,"metrics":{"(Recall@10+Recall@50)/2":"20.42"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-6","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"ImageNet-R","model":"SEARLE-OTI (CLIP B/32)","rank_in_archive_order":13,"of":20,"metrics":{"(Recall@10+Recall@50)/2":"12.77"},"uses_additional_data":false},{"leaderboard":"/sota/zero-shot-composed-image-retrieval-zs-cir-on-6","task":"Zero-Shot Composed Image Retrieval (ZS-CIR)","dataset":"ImageNet-R","model":"SEARLE (CLIP B/32)","rank_in_archive_order":14,"of":20,"metrics":{"(Recall@10+Recall@50)/2":"11.94"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/2303.15247","atlas_url":"https://app.syntology.ai/?focus=2303.15247","mcp":null,"developers":"https://syntology.ai/developers","held":true,"reason":"arxiv_id_in_atlas","checked_against":{"atlas_ids":549449,"atlas_ids_sha256":"afd72399dafd9906e9062927f70655952385a3d75488558e1f0484e950ab09b7","read_at":"2026-10-05T01:46:21Z"},"agent_calls":[{"tool":"get_citation_path","arguments":{"paper_1":"2303.15247"},"arguments_in":{"paper_2":"another paper's arXiv id or title"},"call":"get_citation_path(paper_1=\"2303.15247\", paper_2=\"…\")"},{"tool":"get_concepts_for_paper","arguments":{"arxiv_id":"2303.15247"},"call":"get_concepts_for_paper(arxiv_id=\"2303.15247\")"}],"by_repo":[{"repo":"miccunifi/circo","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0},{"repo":"miccunifi/searle","kind":"official","samples":0,"ran":0,"instrument":0,"unverified":0,"harvested_for_other_papers":1,"harvested_for":["2405.02951"],"harvested_for_calls":[{"arxiv_id":"2405.02951","tool":"get_harvested_code_for_paper","arguments":{"arxiv_id":"2405.02951"},"call":"get_harvested_code_for_paper(arxiv_id=\"2405.02951\")"}]}],"by_repo_harvested_for_is":"harvested_for: the arXiv ids (as Syntology's graph holds them) of every paper whose samples Syntology's graph links to this repository, in arXiv-id order, complete; the page names the first 3 and says how many more. A link only: it does not say the papers are the same work, that one copied the other, or that the code implements this paper's method; harvested_for_calls gives the get_harvested_code_for_paper call for each arXiv-shaped id among them","by_repo_is":"one row per repository in the archive's code links, Syntology's graph links or the samples; samples 0 means no sample is linked to this paper, a statement about Syntology's coverage, not about the repository: with harvested_for_other_papers present, Syntology harvested the repository for that many other papers, and without it, Syntology harvested nothing from it; placed_by_identical_code counts samples whose link records no repository, placed here because identical code was harvested from this repository for another paper; ran counts executions on a synthesized input and instrument counts failures of Syntology's instrument, not of the code"},"code_links_note":null,"arxiv_metadata":null,"syntology_extracted_results":null,"record_sha256":"c92458fa2a7012337513c4b6d1477d9486ef6405ddf6b2a4024248a1647bbd79","record_changed_at":"2026-09-28","record_changed_at_basis":"first_hashed"}