{"about":{"site":"https://codewithpapers.app","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.","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"},"url":"/paper/are-we-done-with-object-recognition-the-icub","title":"Are we done with object recognition? The iCub robot's perspective","arxiv_id":"1709.09882","date":"2017-09-28","proceeding":null,"authors":["Giulia Pasquale","Carlo Ciliberto","Francesca Odone","Lorenzo Rosasco","Lorenzo Natale"],"abstract":"We report on an extensive study of the benefits and limitations of current\ndeep learning approaches to object recognition in robot vision scenarios,\nintroducing a novel dataset used for our investigation. To avoid the biases in\ncurrently available datasets, we consider a natural human-robot interaction\nsetting to design a data-acquisition protocol for visual object recognition on\nthe iCub humanoid robot. Analyzing the performance of off-the-shelf models\ntrained off-line on large-scale image retrieval datasets, we show the necessity\nfor knowledge transfer. We evaluate different ways in which this last step can\nbe done, and identify the major bottlenecks affecting robotic scenarios. By\nstudying both object categorization and identification problems, we highlight\nkey differences between object recognition in robotics applications and in\nimage retrieval tasks, for which the considered deep learning approaches have\nbeen originally designed. In a nutshell, our results confirm the remarkable\nimprovements yield by deep learning in this setting, while pointing to specific\nopen challenges that need be addressed for seamless deployment in robotics.","url_abs":"http://arxiv.org/abs/1709.09882v2","url_pdf":"http://arxiv.org/pdf/1709.09882v2.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":"are-we-done-with-object-recognition-the-icub","repo_url":"https://github.com/VadymV/ICOS","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"deep-learning","task_name":"Deep Learning"},{"task_slug":"image-retrieval","task_name":"Image Retrieval"},{"task_slug":"object","task_name":"Object"},{"task_slug":"object-categorization","task_name":"Object Categorization"},{"task_slug":"object-recognition","task_name":"Object Recognition"},{"task_slug":"retrieval","task_name":"Retrieval"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}