Meta tags:

Headings (most frequently used words):

the, person, on, to, table, as, and, left, his, with, scene, results, in, around, arms, you, motion, generation, affordance, novel, is, forward, then, right, man, from, side, hand, down, up, sit, walk, floor, an, move, say, interact, can, language, guided, human, humanml3d, abstract, humanise, our, evaluation, set, bibtex, walks, clockwise, circle, walking, back, other, direction, jogs, semi, circles, squats, deeply, three, times, while, raising, both, air, if, holding, dumbell, jumps, waves, lie, stand, toilet, refrigerator, chair, wanders, room, dances, bed, happily, someone, stretches, overhead, puts, something, lies, picks, object, off, moving, we, introduce, two, stage, framework, that, employs, intermediate, representation, effectively, linking, 3d, grounding, conditional,

Text of the page (most frequently used words):
the (25), and (21), #motion (11), person (10), scene (9), #affordance (8), language (6), human (6), with (6), our (6), from (5), table (5), for (4), you (4), generation (4), yixin (4), zhang (4), his (4), left (4), page (3), zan (3), wang (3), guided (3), computer (3), arms (3), around (3), results (3), novel (3), generating (3), framework (3), model (3), that (3), technology (3), institute (3), university (3), was (2), project (2), move (2), say (2), interact (2), can (2), author (2), chen (2), jia (2), baoxiong (2), puhao (2), jinlu (2), jingze (2), liu (2), tengyu (2), zhu (2), liang (2), wei (2), huang (2), siyuan (2), cvpr (2), 2024 (2), floor (2), hand (2), down (2), man (2), evaluation (2), set (2), walk (2), sit (2), humanise (2), side (2), right (2), forward (2), then (2), humanml3d (2), challenges (2), these (2), generative (2), models (2), scenes (2), data (2), introduce (2), effectively (2), linking (2), grounding (2), conditional (2), diffusion (2), extensive (2), two (2), stage (2), employs (2), intermediate (2), representation (2), school (2), science (2), beijing (2), peking (2), this, borrowed, which, adopted, please, use, best, experience, send, feedback, questions, chrome, nerfies, academic, template, inproceedings, wang2024move, title, booktitle, proceedings, ieee, cvf, conference, vision, pattern, recognition, year, bibtex, moving, picks, object, off, lies, puts, something, someone, stretches, overhead, dances, bed, happily, wanders, room, chair, refrigerator, toilet, stand, lie, waves, jumps, squats, deeply, three, times, while, raising, both, air, holding, dumbell, jogs, semi, circles, walking, back, other, direction, walks, clockwise, circle, despite, significant, advancements, text, synthesis, within, environments, poses, substantial, stem, primarily, absence, powerful, capable, jointly, modeling, natural, intensive, requirements, contrasted, scarcity, comprehensive, high, quality, datasets, tackle, issues, comprises, adm, predicting, explicit, map, amdm, plausible, motions, leveraging, maps, method, overcomes, difficulty, under, multimodal, condition, signals, especially, when, training, limited, lacking, pairs, experiments, demonstrate, approach, consistently, outperforms, all, baselines, established, benchmarks, including, additionally, validate, exceptional, generalization, capabilities, specially, curated, featuring, previously, unseen, descriptions, abstract, code, video, supplementary, paper, arxiv, indicates, corresponding, national, key, laboratory, general, artificial, intelligence, bigai, dept, automation, tsinghua, cfcs, yangtze, delta, region, academy, jiaxing, highlight, afford, moiton,

Text of the page (random words):
afford moiton move as you say interact as you can language guided human motion generation with scene affordance cvpr 2024 highlight zan wang 1 2 yixin chen 2 baoxiong jia 2 puhao li 2 3 jinlu zhang 2 4 jingze zhang 2 3 tengyu liu 2 yixin zhu 5 ️ wei liang 1 6 ️ siyuan huang 2 ️ 1 school of computer science technology beijing institute of technology 2 national key laboratory of general artificial intelligence bigai 3 dept of automation tsinghua university 4 cfcs school of computer science peking university 5 institute for ai peking university 6 yangtze delta region academy of beijing institute of technology jiaxing ️ indicates corresponding author arxiv paper supplementary video code ️ we introduce a novel two stage framework that employs scene affordance as an intermediate representation effectively linking 3d scene grounding and conditional motion generation abstract despite significant advancements in text to motion synthesis generating language guided human motion within 3d environments poses substantial challenges these challenges stem primarily from i the absence of powerful generative models capable of jointly modeling natural language 3d scenes and human motion and ii the generative models intensive data requirements contrasted with the scarcity of comprehensive high quality language scene motion datasets to tackle these issues we introduce a novel two stage framework that employs scene affordance as an intermediate representation effectively linking 3d scene grounding and conditional motion generation our framework comprises an affordance diffusion model adm for predicting explicit affordance map and an affordance to motion diffusion model amdm for generating plausible human motions by leveraging scene affordance maps our method overcomes the difficulty in generating human motion under multimodal condition signals especially when training with limited data lacking extensive language scene motion pairs our extensive experiments demonstrate that our approach consistently outperforms all baselines on established benchmarks including humanml3d and humanise additionally we validate our model s exceptional generalization capabilities on a specially curated evaluation set featuring previously unseen descriptions and scenes results on humanml3d the person walks in a clockwise circle the person is walking forward and then back the other direction a person jogs forward and semi circles around to the left and then to the right a man squats deeply three times while raising both arms in the air as if holding a dumbell a person jumps from side to side right to left a person waves with his left hand results on humanise lie down on the table stand up from the table sit on the toilet sit on the table walk to the refrigerator walk to the chair results on our novel evaluation set a person wanders in the room around the table a man dances on the bed happily someone stretches his arms overhead a person puts something on the table a person lies down on the floor a person picks an object up off the floor with his left hand a person is moving his arms around bibtex inproceedings wang2024move title move as you say interact as you can language guided human motion generation with scene affordance author wang zan and chen yixin and jia baoxiong and li puhao and zhang jinlu and zhang jingze and liu tengyu and zhu yixin and liang wei and huang siyuan booktitle proceedings of the ieee cvf conference on computer vision and pattern recognition cvpr year 2024 this page was borrowed from the academic project page template which was adopted from the nerfies project page please use chrome for best experience send feedback and questions to zan wang