[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-141364-en":3,"doc-seo-141364-105":30,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},141364,2336475104042,"Tawan","https://ap-avatar.wpscdn.com/avatar/22000c4c32af1715be0?x-image-process=image/resize,m_fixed,w_180,h_180&k=1786537525561427321",6,"Technology","Learning Any-View 6DoF Robotic Grasping in Cluttered Scenes via Neural Surface Rendering - Robotic grasp detection abstract","A significant challenge for real-world robotic manipulation is performing effective 6DoF grasping of objects in cluttered scenes from a single viewpoint, without further scene exploration. The work reinterprets grasping as rendering and introduces NeuGraspNet, which detects 6DoF grasps using neural volumetric representations and surface rendering. It encodes the robot end-effector and object surface interaction by jointly learning to render local object surfaces and grasping functions in a shared feature space. Global scene-level features generate grasp candidates, while local neural surface features evaluate grasp quality. The method enables fully implicit 6DoF grasp quality prediction in partially observed scenes, operates on random viewpoints common to mobile manipulation, and outperforms existing implicit and semi-implicit approaches, validated with a mobile manipulator in real-world settings.","Robotics: Science and Systems 2024  \nDelft, Netherlands, July 15-July 19, 2024  \nLearning Any-View 6DoF Robotic Grasping in Cluttered Scenes via Neural Surface Rendering  \nSnehal Jauhri 1 , Ishikaa Lunawat2 , and Georgia Chalvatzaki 1 ,3 ,4  \n1 Computer Science Dept., TU Darmstadt, Germany  \n2 NIT Trichy, India 3 Hessian.AI, Darmstadt, Germany  \n4 Center for Mind, Brain and Behavior, Uni. Marburg and JLU Giessen, Germany  \nFig. 1: Grasping as rendering. Our network, NeuGraspNet, uses a single random-view depth input, encodes the scene in an implicit feature volume, and uses multi-level rendering to select relevant features and predict grasping functions. As shown, NeuGraspNet generalizes to random-view mobile manipulation grasping scenarios. More at: [sites.google.com/view/neugraspnet](sites.google.com/view/neugraspnet)  \nAbstract—A significant challenge for real-world robotic manipulation is the effective 6DoF grasping of objects in cluttered scenes from any single viewpoint without needing additional scene exploration. This work re-interprets grasping as rendering and introduces NeuGraspNet, a novel method for 6DoF grasp detection that leverages advances in neural volumetric representationsand surface rendering. We encode the interaction between a robot’s end-effector and an object’s surface by jointly learning to render the local object surface and learning grasping functions in a shared feature space. Our approach uses global (scenelevel) features for grasp generation and local (grasp-level) neural surface features for grasp evaluation. This enables effective, fully implicit 6DoF grasp quality prediction, even in partially observed scenes. NeuGraspNet operates on random viewpoints, common in mobile manipulation scenarios, and outperforms existing implicit and semi-implicit grasping methods. We demonstrate the realworld applicability of the method with a mobile manipulator robot, grasping in open cluttered spaces.  \nI. INTRODUCTION  \nRobotic manipulation is crucial for enabling various applications such as home-assistance, industrial automation etc. A key component for manipulation is the ability to grasp objects in unstructured, cluttered spaces under partial observability. This ability would enhance the efficiency, versatility, and autonomy of robots operating in everyday environments. Deep learning has been crucial in making advances in robotic grasping [37, 38, 44, 9] by training networks using simulation data and transferring to the real world. However, 6DoF grasping in the wild, i.e., grasping in the SE(3) space from any viewpoint  \nremains a challenge [32, 47, 53] . Embodied AI agents, e.g., mobile manipulation robots [27, 25], are expected to perform manipulation tasks similar to humans; humans can leverage geometric information from limited views and mental object models to grasp objects without exploring the whole scene. Such grasping in open cluttered spaces requires that robots, given some spatial information, e.g., 3D pointcloud data, can reconstruct the scene, understand graspable areas of objects, and detect grasps likely to succeed. Moreover, robots should reason about the grasp’s affordance [21], i.e., the subsequent task a grasp allows, while additionally avoiding collision with the surrounding environment.  \n6DoF grasping methods can be classified into methods that explicitly generate grasp poses [45, 61, 58] or implicitly classify the grasp quality of any grasp candidate in SE(3) using a discriminative model [59, 34, 26] . The ability to assess the quality of any grasp pose implicitly is essential to applications in which grasp candidates are pre-defined due to human demonstrations [67] or other affordance-based information [15, 33] . Moreover, explicit generative models are difficult to combine with additional constraints since the constraints can only be applied as a post-filtering step. In implicit methods, however, the distribution of grasp candidates can be chosen and constrained before querying the","cbCaioGuuzYbJSrU","https://ap.wps.com/l/cbCaioGuuzYbJSrU","pdf",3958384,1,12,"English","en",105,"# Abstract\n## Challenge: 6DoF grasping from any single view\n## Approach: grasping as rendering with NeuGraspNet\n## Method details: global generation and local evaluation\n## Evaluation and real-world applicability\n# I. INTRODUCTION\n## Robotic manipulation and grasping needs\n## Explicit vs implicit 6DoF grasp methods\n## Neural scene representations and partial observability","[{\"question\":\"What problem does NeuGraspNet address?\",\"answer\":\"It addresses effective 6DoF grasping of objects in cluttered scenes from any single viewpoint without needing additional scene exploration.\"},{\"question\":\"How does NeuGraspNet model grasping?\",\"answer\":\"It reinterprets grasping as rendering, jointly learning to render the local object surface and grasping functions in a shared feature space.\"},{\"question\":\"What is the difference between global and local features in the method?\",\"answer\":\"Global (scene-level) features are used for grasp generation, while local (grasp-level) neural surface features are used for grasp evaluation and quality prediction.\"}]","Learning Any-View 6DoF Robotic Grasping in Cluttered Scenes via Neural Surface Rendering - Robotic grasp detection abstract | PDF",1787654930,30,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":28},"learning-any-view-6dof-robotic-grasping-in-cluttered-scenes-via-neural-surface-rendering-robotic-grasp-detection-abstract","",{"@graph":36,"@context":86},[37,54,69],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/technology/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/learning-any-view-6dof-robotic-grasping-in-cluttered-scenes-via-neural-surface-rendering-robotic-grasp-detection-abstract/141364/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-09-04","2026-08-25",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"What problem does NeuGraspNet address?","Question",{"text":76,"@type":77},"It addresses effective 6DoF grasping of objects in cluttered scenes from any single viewpoint without needing additional scene exploration.","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"How does NeuGraspNet model grasping?",{"text":81,"@type":77},"It reinterprets grasping as rendering, jointly learning to render the local object surface and grasping functions in a shared feature space.",{"name":83,"@type":74,"acceptedAnswer":84},"What is the difference between global and local features in the method?",{"text":85,"@type":77},"Global (scene-level) features are used for grasp generation, while local (grasp-level) neural surface features are used for grasp evaluation and quality prediction.","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":93},[94,98,102,106,111,114,119,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":46,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":112,"slug":113},50,"technology",{"id":115,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":120,"doc_module":4,"doc_module_name":46,"category_name":121,"show_sort_weight":29,"slug":122},8,"Research & Report","research-report",{"id":124,"doc_module":4,"doc_module_name":46,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":46,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":46,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":46,"category_name":137,"show_sort_weight":107,"slug":138},19,"General","general"]