--- title: Competing in RoboCup@Home and winning the 2026 best... description: A team from the University of Bonn won the best paper award at RoboCup 2026 for a method that semi-automatically labels objects robots must recognize during... canonical: https://daily.dev/posts/competing-in-robocup-home-and-winning-the-2026-best-paper-award-an-interview-with-raphael-memmeshei-ercwsug6k twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Competing in RoboCup@Home and winning the 2026 best paper award: an interview with Raphael Memmesheimer | daily.dev og:description: A team from the University of Bonn won the best paper award at RoboCup 2026 for a method that semi-automatically labels objects robots must recognize during... og:url: https://daily.dev/posts/competing-in-robocup-home-and-winning-the-2026-best-paper-award-an-interview-with-raphael-memmeshei-ercwsug6k og:image: https://api.daily.dev/og/posts/ercWsug6k.png og:image:alt: Competing in RoboCup@Home and winning the 2026 best paper award: an interview with Raphael Memmesheimer og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Competing in RoboCup@Home and winning the 2026 best paper award: an interview with Raphael Memmesheimer **[Robohub](https://daily.dev/sources/robohub)** · 10 min read · 0 upvotes · 0 comments ## Summary A team from the University of Bonn won the best paper award at RoboCup 2026 for a method that semi-automatically labels objects robots must recognize during competitions. The approach segments objects, then uses reference images to semi-automatically label them, automatically labeling around 60% of the data. This matters because teams get only two preparation days to teach robots up to 50 unfamiliar objects per competition. The team, NimbRo led by Raphael Memmesheimer, placed 3rd in this year's World Cup in Korea but performed best in the new 'Doing Laundry' task, where its robot opened a washing machine, pulled out t-shirts by feel, and folded them. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How does semi-supervised label propagation reduce manual image annotation for robot vision training? It combines a generic segmentation model that flags 'something that might be an object' with reference images of actual competition objects, then propagates labels automatically. This approach automatically labels around 60 percent of the data, letting a robotics team quickly build vision models for up to 50 unseen objects within a two-day competition prep window instead of relying on slow manual annotation. _Developers building vision pipelines can track efficient annotation techniques like this one on daily.dev._ ### Why don't open-vocabulary or promptable object detection models work well for recognizing many similar objects like drinks or snacks? They struggle to scale past small class counts, working fine with 5 to 10 classes but failing to distinguish between roughly 50 object classes needed in robot competitions, especially within categories like drinks. Teams found these methods have particular trouble telling apart objects of the same category, such as different drink brands, which pushed them toward a custom segmentation and label-propagation pipeline instead. _Anyone evaluating open-vocabulary detection against custom pipelines can follow real-world comparisons on daily.dev._ ### What tasks does a robot have to complete in the RoboCup@Home competition? Competing robots handle five main tasks: human-robot interaction at a party (greeting and seating guests), a pick-and-place table-clearing task involving a dishwasher, a general-purpose service task driven by random natural language commands, a restaurant task in an unknown, noisy location, and a laundry task requiring opening a washing machine, removing t-shirts by feel, and folding them. _Robotics engineers tracking competition benchmarks like these can follow updates on daily.dev._ ## Similar posts on daily.dev - [Video Friday: A World Cup for Robots](https://daily.dev/posts/video-friday-a-world-cup-for-robots-olnbgngmi) · IEEE Spectrum · 0 upvotes · 0 comments - [A history of RoboCup with Manuela Veloso](https://daily.dev/posts/a-history-of-robocup-with-manuela-veloso-kkcnfmehp) · Robohub · 0 upvotes · 0 comments - [Powered by Clearpath: Research Highlights from ICRA 2026](https://daily.dev/posts/powered-by-clearpath-research-highlights-from-icra-2026-xuxreruva) · Clearpath Robotics · 0 upvotes · 0 comments - [Daniela Rus receives Bavarian Minister-President's High-Tech Prize](https://daily.dev/posts/daniela-rus-receives-bavarian-minister-president-s-high-tech-prize-xffdzl3iq) · MIT News · 0 upvotes · 0 comments --- Tags: [#robotics](https://daily.dev/tags/robotics), [#computer-vision](https://daily.dev/tags/computer-vision) [View this post on daily.dev](https://daily.dev/posts/competing-in-robocup-home-and-winning-the-2026-best-paper-award-an-interview-with-raphael-memmeshei-ercwsug6k) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Competing in RoboCup@Home and winning the 2026 best paper award: an interview with Raphael Memmesheimer","url":"https://daily.dev/posts/competing-in-robocup-home-and-winning-the-2026-best-paper-award-an-interview-with-raphael-memmeshei-ercwsug6k","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/competing-in-robocup-home-and-winning-the-2026-best-paper-award-an-interview-with-raphael-memmeshei-ercwsug6k"},"datePublished":"2026-10-28T09:16:55.451Z","dateModified":"2026-10-28T09:17:44.836Z","description":"A team from the University of Bonn won the best paper award at RoboCup 2026 for a method that semi-automatically labels objects robots must recognize during...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/791a9f0d49cc2875150e25f049cbfbb1?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/791a9f0d49cc2875150e25f049cbfbb1?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Robohub","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Robohub","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/79911e4f3a514623a70a98793133f536","url":"https://daily.dev/sources/robohub"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/competing-in-robocup-home-and-winning-the-2026-best-paper-award-an-interview-with-raphael-memmeshei-ercwsug6k","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"robotics,computer-vision","timeRequired":"PT10M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Robohub","item":"https://daily.dev/sources/robohub"},{"@type":"ListItem","position":3,"name":"Competing in RoboCup@Home and winning the 2026 best paper award: an interview with Raphael Memmesheimer"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/competing-in-robocup-home-and-winning-the-2026-best-paper-award-an-interview-with-raphael-memmeshei-ercwsug6k#faq","mainEntity":[{"@type":"Question","name":"How does semi-supervised label propagation reduce manual image annotation for robot vision training?","acceptedAnswer":{"@type":"Answer","text":"It combines a generic segmentation model that flags 'something that might be an object' with reference images of actual competition objects, then propagates labels automatically. This approach automatically labels around 60 percent of the data, letting a robotics team quickly build vision models for up to 50 unseen objects within a two-day competition prep window instead of relying on slow manual annotation. Developers building vision pipelines can track efficient annotation techniques like this one on daily.dev."}},{"@type":"Question","name":"Why don't open-vocabulary or promptable object detection models work well for recognizing many similar objects like drinks or snacks?","acceptedAnswer":{"@type":"Answer","text":"They struggle to scale past small class counts, working fine with 5 to 10 classes but failing to distinguish between roughly 50 object classes needed in robot competitions, especially within categories like drinks. Teams found these methods have particular trouble telling apart objects of the same category, such as different drink brands, which pushed them toward a custom segmentation and label-propagation pipeline instead. Anyone evaluating open-vocabulary detection against custom pipelines can follow real-world comparisons on daily.dev."}},{"@type":"Question","name":"What tasks does a robot have to complete in the RoboCup@Home competition?","acceptedAnswer":{"@type":"Answer","text":"Competing robots handle five main tasks: human-robot interaction at a party (greeting and seating guests), a pick-and-place table-clearing task involving a dishwasher, a general-purpose service task driven by random natural language commands, a restaurant task in an unknown, noisy location, and a laundry task requiring opening a washing machine, removing t-shirts by feel, and folding them. Robotics engineers tracking competition benchmarks like these can follow updates on daily.dev."}}]} ```