Meta tags:
author=Anonymous;

Headings (most frequently used words):

sleeve, metafold, language, guided, multi, category, garment, folding, framework, via, trajectory, generation, and, foundation, model, abstract, pipeline, overview, dataset, visualization, bibtex, accepted, to, iros, 2025, oral, presentation, no, short, long, pants,

Text of the page (most frequently used words):
and (23), action (16), the (14), sleeve (12), folding (10), #garment (7), #trajectory (7), model (7), metafold (6), pants (6), language (5), this (4), for (4), framework (4), generation (4), long (4), short (4), each (4), university (4), guided (3), multi (3), category (3), foundation (3), arxiv (3), from (3), dataset (3), garments (3), stages (3), task (3), via (2), author (2), haonan (2), chen (2), junxiao (2), ruihai (2), yiwei (2), liu (2), yiwen (2), hou (2), zhixuan (2), jingxiang (2), guo (2), chongkai (2), gao (2), zhenyu (2), wei (2), shensi (2), jiaqi (2), huang (2), lin (2), shao (2), 2025 (2), 2503 (2), 08372 (2), nosleeve (2), shortsleeve (2), longsleeve (2), row (2), three (2), process (2), subsequent (2), has (2), generated (2), using (2), point (2), cloud (2), corresponding (2), points (2), enabling (2), overview (2), manipulation (2), paper (2), planning (2), prediction (2), learning (2), website, designed, based, thanks, your, visiting, total, page, views, nerfies, misc, chen2025metafoldlanguageguidedmulticategorygarment, title, year, eprint, archiveprefix, primaryclass, url, https, org, abs, bibtex, present, visualizations, several, trajectories, these, include, contains, different, dragging, slider, beneath, reveals, its, selecting, right, side, visualizes, only, one, stage, have, two, auto, play, visualization, data, clothing, heuristic, methods, diffclothai, simulation, environment, with, descriptions, subsequently, added, takes, any, given, frame, description, inputs, generate, fed, into, manifoundation, estimate, contact, force, directions, robot, conduct, actions, then, iteratively, refined, feedback, loop, blue, orange, green, pipeline, common, yet, challenging, robotic, deformability, leads, vast, state, space, complex, dynamics, complicates, precise, fine, grained, previous, approaches, often, rely, predefined, key, demonstrations, constraining, their, generalizability, across, diverse, categories, presents, that, disentangles, independently, enhance, generalization, employs, low, level, structure, facilitates, adapt, flexibly, various, user, instructions, tasks, experimental, results, demonstrate, our, proposed, superiority, abstract, code, accepted, iros, oral, presentation, national, singapore, nus, guangzhou, research, translation, innovation, institute, nanjing, peking, shanghai, jiao, tong, denotes, equal, contribution,

Text of the page (random words):
metafold metafold language guided multi category garment folding framework via trajectory generation and foundation model haonan chen 1 2 junxiao li 3 ruihai wu 4 yiwei liu 1 yiwen hou 1 zhixuan xu 1 jingxiang guo 1 chongkai gao 1 zhenyu wei 5 shensi xu 3 jiaqi huang 3 lin shao 1 2 1 national university of singapore 2 nus guangzhou research translation and innovation institute 3 nanjing university 4 peking university 5 shanghai jiao tong university denotes equal contribution corresponding author accepted to iros 2025 oral presentation arxiv paper code dataset abstract garment folding is a common yet challenging task in robotic manipulation the deformability of garments leads to a vast state space and complex dynamics and complicates precise fine grained manipulation previous approaches often rely on predefined key points or demonstrations constraining their generalizability across diverse garment categories this paper presents a framework metafold that disentangles task planning from action prediction learning each independently to enhance model generalization it employs language guided point cloud trajectory generation for task planning and a low level foundation model for action prediction this structure facilitates multi category learning enabling the model to adapt flexibly to various user instructions and folding tasks experimental results demonstrate our proposed framework s superiority pipeline overview overview the folding trajectory data for clothing is generated using heuristic methods in the diffclothai simulation environment with language descriptions subsequently added green the trajectory generation model takes a point cloud from any given frame and a corresponding language description as inputs to generate the subsequent trajectory orange the generated trajectory is fed into the manifoundation model to estimate contact points and force directions enabling the robot to conduct garment folding actions this process is then iteratively refined using a feedback loop blue metafold dataset visualization auto play we present visualizations of several garment folding trajectories from the metafold dataset these garments include nosleeve shortsleeve longsleeve and pants each row contains three different garments dragging the slider beneath a garment reveals its folding process selecting the action on the right side of each row visualizes the subsequent folding stages nosleeve has only one folding stage shortsleeve and longsleeve each have three stages and pants has two stages no sleeve no sleeve 1 00 20 no sleeve 2 00 20 no sleeve 3 00 20 action action 0 short sleeve short sleeve 1 00 20 short sleeve 2 00 20 short sleeve 3 00 20 action action 0 action 1 action 2 long sleeve long sleeve 1 00 20 long sleeve 2 00 20 long sleeve 3 00 20 action action 0 action 1 action 2 pants pants 1 00 20 pants 2 00 20 pants 3 00 20 action action 0 action 1 bibtex misc chen2025metafoldlanguageguidedmulticategorygarment title metafold language guided multi category garment folding framework via trajectory generation and foundation model author haonan chen and junxiao li and ruihai wu and yiwei liu and yiwen hou and zhixuan xu and jingxiang guo and chongkai gao and zhenyu wei and shensi xu and jiaqi huang and lin shao year 2025 eprint 2503 08372 archiveprefix arxiv primaryclass cs ro url https arxiv org abs 2503 08372 this website is designed based on nerfies thanks for your visiting total page views