<span style=\"color: #f14668;\">BMVC 2026</span></b>)\n\nIn this work, we present:\n\n* **`QCell`**, a novel query-based model for overlapping cell instance segmentation ⭐️\n* A new overlapping object segmentation dataset: **`Organoids`** 🔥\n\n🔗 **GitHub**: https://github.com/SlavkoPrytula/QCell\n🌐 **Project page**: https://slavkoprytula.github.io/QCell/\n\nIf you find this work useful, consider giving it a ⭐️ on GitHub to support further open-source research!","html":"<h1 class=\"relative group flex items-baseline\">\n\t<a id=\"qcell-query-based-cell-instance-segmentation-bmvc-2026\" class=\"block pr-1.5 text-lg md:absolute md:p-1.5 md:opacity-0 md:group-hover:opacity-100 md:right-full\" href=\"#qcell-query-based-cell-instance-segmentation-bmvc-2026\" rel=\"nofollow\">\n\t\t<span class=\"header-link\"><svg class=\"text-gray-500 hover:text-black dark:hover:text-gray-200 w-4\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\" aria-hidden=\"true\" role=\"img\" width=\"1em\" height=\"1em\" preserveAspectRatio=\"xMidYMid meet\" viewBox=\"0 0 256 256\"><path d=\"M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z\" fill=\"currentColor\"></path></svg></span>\n\t</a>\n\t<span>\n\t\tQCell: Query-Based Cell Instance Segmentation (<b><span style=\"color: #f14668;\">BMVC 2026</span></b>)\n\t</span>\n</h1>\n<p>In this work, we present:</p>\n<ul>\n<li><strong><code>QCell</code></strong>, a novel query-based model for overlapping cell instance segmentation ⭐️</li>\n<li>A new overlapping object segmentation dataset: <strong><code>Organoids</code></strong> 🔥</li>\n</ul>\n<p>🔗 <strong>GitHub</strong>: <a href=\"https://github.com/SlavkoPrytula/QCell\" rel=\"nofollow\">https://github.com/SlavkoPrytula/QCell</a><br>🌐 <strong>Project page</strong>: <a href=\"https://slavkoprytula.github.io/QCell/\" rel=\"nofollow\">https://slavkoprytula.github.io/QCell/</a></p>\n<p>If you find this work useful, consider giving it a ⭐️ on GitHub to support further open-source research!</p>\n","updatedAt":"2026-09-04T07:08:48.111Z","author":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","fullname":"Yaroslav Prytula","name":"YaroslavPrytula","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false}},"numEdits":4,"identifiedLanguage":{"language":"en","probability":0.8609561920166016},"editors":["YaroslavPrytula"],"editorAvatarUrls":["/avatars/28871a0725cf49347457fa8535ce5fdb.svg"],"reactions":[{"reaction":"❤️","users":["YaroslavPrytula"],"count":1}],"isReport":false}},{"id":"6a9a6d43fec692165b092d82","author":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","fullname":"Yaroslav Prytula","name":"YaroslavPrytula","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false},"createdAt":"2026-09-04T07:03:31.000Z","type":"comment","data":{"edited":true,"hidden":false,"latest":{"raw":"\n\n**Overview of QCell.** QCell builds on a MaskDINO-style query-based segmentation pipeline, where multi-scale image features and positional embeddings are processed by the encoder and refined by transformer decoder layers with content and DN queries. The proposed modules are shown above: (a) **instance recombination** decomposes each query into amodal, visible, and occluded sub-representations and recombines them into a refined full-instance query; (b) **contrastive query learning** uses matched instance queries `q̂_i` by Hungarian matching as anchors, corresponding DN queries `d̂_i^+` across all groups as positives, and other DN queries as negatives to align queries of the same cell and separate queries of different cells in latent space.","html":"<p><a href=\"https://cdn-uploads.huggingface.co/production/uploads/662798e43e2a8eec1b53ed1a/7lpFXZ7z4tfZy2vJUdyta.png\" rel=\"nofollow\"><img src=\"https://cdn-uploads.huggingface.co/production/uploads/662798e43e2a8eec1b53ed1a/7lpFXZ7z4tfZy2vJUdyta.png\" alt=\"qcell\"></a></p>\n<p><strong>Overview of QCell.</strong> QCell builds on a MaskDINO-style query-based segmentation pipeline, where multi-scale image features and positional embeddings are processed by the encoder and refined by transformer decoder layers with content and DN queries. The proposed modules are shown above: (a) <strong>instance recombination</strong> decomposes each query into amodal, visible, and occluded sub-representations and recombines them into a refined full-instance query; (b) <strong>contrastive query learning</strong> uses matched instance queries <code>q̂_i</code> by Hungarian matching as anchors, corresponding DN queries <code>d̂_i^+</code> across all groups as positives, and other DN queries as negatives to align queries of the same cell and separate queries of different cells in latent space.</p>\n","updatedAt":"2026-09-04T07:07:35.606Z","author":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","fullname":"Yaroslav Prytula","name":"YaroslavPrytula","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false}},"numEdits":1,"identifiedLanguage":{"language":"en","probability":0.8261122703552246},"editors":["YaroslavPrytula"],"editorAvatarUrls":["/avatars/28871a0725cf49347457fa8535ce5fdb.svg"],"reactions":[{"reaction":"🚀","users":["YaroslavPrytula"],"count":1}],"isReport":false}},{"id":"6a9a6e3009d3992b54d86452","author":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","fullname":"Yaroslav Prytula","name":"YaroslavPrytula","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false},"createdAt":"2026-09-04T07:07:28.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"\n\n**Organoids.** One of our key contributions is a novel `Organoids` dataset for overlapping object segmentation in brightfield microscopy. The dataset contains 1,186 training images, 1,199 validation images, and 201 test images at a resolution of 540 × 540. The dataset presents dense and highly overlapping scenes, with up to 105 instances per training image and an average of 96 instances per test image, reaching a maximum of 223. This makes Organoids a challenging real-world benchmark for evaluating instance separation and de-overlapping in microscopy. Additional dataset details are provided in the supplementary material.","html":"<p><a href=\"https://cdn-uploads.huggingface.co/production/uploads/662798e43e2a8eec1b53ed1a/mYEQs2HVRYTLXhaO3EBxZ.png\" rel=\"nofollow\"><img src=\"https://cdn-uploads.huggingface.co/production/uploads/662798e43e2a8eec1b53ed1a/mYEQs2HVRYTLXhaO3EBxZ.png\" alt=\"organoids\"></a></p>\n<p><strong>Organoids.</strong> One of our key contributions is a novel <code>Organoids</code> dataset for overlapping object segmentation in brightfield microscopy. The dataset contains 1,186 training images, 1,199 validation images, and 201 test images at a resolution of 540 × 540. The dataset presents dense and highly overlapping scenes, with up to 105 instances per training image and an average of 96 instances per test image, reaching a maximum of 223. This makes Organoids a challenging real-world benchmark for evaluating instance separation and de-overlapping in microscopy. Additional dataset details are provided in the supplementary material.</p>\n","updatedAt":"2026-09-04T07:07:28.150Z","author":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","fullname":"Yaroslav Prytula","name":"YaroslavPrytula","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8261122703552246},"editors":["YaroslavPrytula"],"editorAvatarUrls":["/avatars/28871a0725cf49347457fa8535ce5fdb.svg"],"reactions":[{"reaction":"🔥","users":["YaroslavPrytula"],"count":1}],"isReport":false}},{"id":"6a9ab1aed0d14682e92f350e","author":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","fullname":"Yaroslav Prytula","name":"YaroslavPrytula","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false},"createdAt":"2026-09-04T11:55:26.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"QCell segments overlapping cells using an instance recombination module and contrastive query alignment, achieving state-of-the-art results against prior amodal segmentation methods. We additionally introduce a new Organoid dataset benchmark for overlapping cell segmentation. Accepted at <b><span style=\"color: #f14668;\">BMVC 2026</span></b>.","html":"<p>QCell segments overlapping cells using an instance recombination module and contrastive query alignment, achieving state-of-the-art results against prior amodal segmentation methods. We additionally introduce a new Organoid dataset benchmark for overlapping cell segmentation. Accepted at <b><span style=\"color: #f14668;\">BMVC 2026</span></b>.</p>\n","updatedAt":"2026-09-04T11:55:26.415Z","author":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","fullname":"Yaroslav Prytula","name":"YaroslavPrytula","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":2,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.8197977542877197},"editors":["YaroslavPrytula"],"editorAvatarUrls":["/avatars/28871a0725cf49347457fa8535ce5fdb.svg"],"reactions":[{"reaction":"🔥","users":["YaroslavPrytula"],"count":1}],"isReport":false}},{"id":"6a9b6f0127087a684139052f","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":379,"isUserFollowing":false},"createdAt":"2026-09-05T01:23:13.000Z","type":"comment","data":{"edited":false,"hidden":false,"latest":{"raw":"This is an automated message from the [Librarian Bot](https://huggingface.co/librarian-bots). I found the following papers similar to this paper. \n\nThe following papers were recommended by the Semantic Scholar API \n\n* [QueenVIS: Rethinking Image-Only Training for Video Instance Segmentation via Query Enrichment](https://huggingface.co/papers/2607.24598) (2026)\n* [InstancePin: Instance-Addressable Layout-to-Image Diffusion via Coordinate Pinning](https://huggingface.co/papers/2608.00588) (2026)\n* [SOS! : A Streamlined Object-Conditional Transformer for Model-free Segmentation](https://huggingface.co/papers/2608.15295) (2026)\n* [`Attention-Guided Cross-Temporal Clustering for Self-Supervised Video Object Segmentation](https://huggingface.co/papers/2607.07230) (2026)\n* [OBBSeg: Irregular Lesion Segmentation under Oriented Bounding Box Annotations](https://huggingface.co/papers/2607.06007) (2026)\n* [Decoupling Language Guidance from Backbones for Text-Guided Medical Segmentation](https://huggingface.co/papers/2607.09481) (2026)\n* [Unified Video Dense Prediction from Disjoint Data](https://huggingface.co/papers/2607.21592) (2026)\n\n\n Please give a thumbs up to this comment if you found it helpful!\n\n If you want recommendations for any Paper on Hugging Face checkout [this](https://huggingface.co/spaces/librarian-bots/recommend_similar_papers) Space\n\n You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: `@librarian-bot recommend`","html":"<p>This is an automated message from the <a href=\"https://huggingface.co/librarian-bots\">Librarian Bot</a>. I found the following papers similar to this paper. </p>\n<p>The following papers were recommended by the Semantic Scholar API </p>\n<ul>\n<li><a href=\"https://huggingface.co/papers/2607.24598\">QueenVIS: Rethinking Image-Only Training for Video Instance Segmentation via Query Enrichment</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2608.00588\">InstancePin: Instance-Addressable Layout-to-Image Diffusion via Coordinate Pinning</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2608.15295\">SOS! : A Streamlined Object-Conditional Transformer for Model-free Segmentation</a> (2026)</li>\n<li>[`Attention-Guided Cross-Temporal Clustering for Self-Supervised Video Object Segmentation](<a href=\"https://huggingface.co/papers/2607.07230\">https://huggingface.co/papers/2607.07230</a>) (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.06007\">OBBSeg: Irregular Lesion Segmentation under Oriented Bounding Box Annotations</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.09481\">Decoupling Language Guidance from Backbones for Text-Guided Medical Segmentation</a> (2026)</li>\n<li><a href=\"https://huggingface.co/papers/2607.21592\">Unified Video Dense Prediction from Disjoint Data</a> (2026)</li>\n</ul>\n<p> Please give a thumbs up to this comment if you found it helpful!</p>\n<p> If you want recommendations for any Paper on Hugging Face checkout <a href=\"https://huggingface.co/spaces/librarian-bots/recommend_similar_papers\">this</a> Space</p>\n<p> You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: <code>@librarian-bot recommend</code></p>\n","updatedAt":"2026-09-05T01:23:13.400Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":379,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.7059404850006104},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.29253","authors":[{"_id":"6a993130fea81827432200c1","user":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","isPro":false,"fullname":"Yaroslav Prytula","user":"YaroslavPrytula","type":"user","name":"YaroslavPrytula"},"name":"Yaroslav Prytula","status":"claimed_verified","statusLastChangedAt":"2026-09-03T08:45:04.426Z","hidden":false},{"_id":"6a993130fea81827432200c2","name":"Anton Popov","hidden":false},{"_id":"6a993130fea81827432200c3","name":"Dmytro Fishman","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/662798e43e2a8eec1b53ed1a/K0UKj6HGsJrJOsO1C99os.png","https://cdn-uploads.huggingface.co/production/uploads/662798e43e2a8eec1b53ed1a/M-9Wnu3AikDVfKnzuH19P.png"],"publishedAt":"2026-08-29T00:00:00.000Z","submittedOnDailyAt":"2026-09-04T00:00:00.000Z","title":"QCell: Recombining and Aligning Cell Queries for Overlapping Instance Segmentation","submittedOnDailyBy":{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","isPro":false,"fullname":"Yaroslav Prytula","user":"YaroslavPrytula","type":"user","name":"YaroslavPrytula"},"summary":"Instance segmentation of overlapping cells in microscopy remains challenging due to semi-transparent structures that produce weak boundaries and mixed visual evidence in overlap regions. Existing methods address this through local regions of interest or shape priors but lack global reasoning across overlapping objects. We present QCell, a novel query-based model that de-overlaps cell instances in microscopy scenes. Our approach combines (i) an instance recombination module that decomposes and recombines query representations in latent space, enabling the model to reason about complete object structure under overlap, and (ii) a contrastive query alignment objective that combines distinctive instance feature learning and separation of overlapping cell queries. We additionally introduce a new Organoid dataset benchmark for overlapping cell segmentation. We show that QCell outperforms state-of-the-art methods across multiple benchmarks, achieving +2.2 AP and +2.7 AJI on ISBI2014. Code is available at https://github.com/SlavkoPrytula/QCell","upvotes":13,"discussionId":"6a993130fea81827432200c4","projectPage":"https://slavkoprytula.github.io/QCell/","githubRepo":"https://github.com/SlavkoPrytula/QCell","githubRepoAddedBy":"user","ai_summary":"QCell is a query-based model that improves instance segmentation of overlapping microscopy cells through latent-space recombination and contrastive query alignment.","ai_keywords":["query-based model","instance recombination","contrastive query alignment","overlapping cell segmentation","microscopy"],"ai_summary_model":"thinkingmachines/Inkling-Small","githubStars":9},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"662798e43e2a8eec1b53ed1a","avatarUrl":"/avatars/28871a0725cf49347457fa8535ce5fdb.svg","isPro":false,"fullname":"Yaroslav Prytula","user":"YaroslavPrytula","type":"user"},{"_id":"654cd3283a321f06c2b66521","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/654cd3283a321f06c2b66521/CT8NhmLXNF5IsWaJoXnr8.jpeg","isPro":false,"fullname":"Oleksandr Kosovan","user":"oleksa-kosovan","type":"user"},{"_id":"6a6a94c6c2dfbd72646cfca2","avatarUrl":"/avatars/458eefceeb99b324b72955c0f71d3671.svg","isPro":false,"fullname":"Daniel Williams","user":"GraniteNoah","type":"user"},{"_id":"6a6c8532da65172f47ef3f9d","avatarUrl":"/avatars/4c207375dd1c09ba60f2cc666d6b01eb.svg","isPro":false,"fullname":"Brian Williams","user":"EmberGlade","type":"user"},{"_id":"6a6c88165f98c1c73ae06866","avatarUrl":"/avatars/71e8dfcd3d7865aef0e4a9fa23e81582.svg","isPro":false,"fullname":"Sarah Brown","user":"sarah-brown","type":"user"},{"_id":"6a6deb9bf372a517696cd7bf","avatarUrl":"/avatars/60b96729136731690470a9d8decac041.svg","isPro":false,"fullname":"James White","user":"HarborJames","type":"user"},{"_id":"6a701daa6786e1fd400119eb","avatarUrl":"/avatars/c5f4fcb614f533438b5976a70bcbae2b.svg","isPro":false,"fullname":"James Lee","user":"Lunar-Bloom","type":"user"},{"_id":"6a6a934c8acf46140bae7c85","avatarUrl":"/avatars/73c414a228d20b3c13a638a534375414.svg","isPro":false,"fullname":"Matthew Thompson","user":"matthew-thompson","type":"user"},{"_id":"6a6aa3fbb58832f7d0fd7e32","avatarUrl":"/avatars/e2b99552157ef3356352ec0af387396e.svg","isPro":false,"fullname":"Sarah Smith","user":"cobaltEvan","type":"user"},{"_id":"6a6de53ec51edbf08f1121ac","avatarUrl":"/avatars/8bcda53af64042ad674b08dd754ab666.svg","isPro":false,"fullname":"Karen Smith","user":"KarenSmith","type":"user"},{"_id":"6a7e7e1df53df3a86dd1aedd","avatarUrl":"/avatars/f9c7ae1731762fa39d126b225bcc2dba.svg","isPro":false,"fullname":"Jordan Smith","user":"rapidarc","type":"user"},{"_id":"6a9ab85ed49d89aa717e70bb","avatarUrl":"/avatars/e1546a3e0a8185728186b1cc95206e37.svg","isPro":false,"fullname":"Melissa Charles","user":"wolferussell14","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.29253.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2608.29253 in a model README.md to link it from this page.

Cite arxiv.org/abs/2608.29253 in a Space README.md to link it from this page.