Meta tags:
author= Despoina Paschalidou;
description= Despoina Paschalidou is a Senior Research Scientist at the NVIDIA Spatial Intelligence Lab, working on computer vision and graphics: 3D reconstruction, generative models, and embodied intelligence.;
Headings (most frequently used words):
-
Text of the page (most frequently used words):
and (205), the (102), for (47), that (41), model (40), our (38), with (31), paschalidou (27), despoina (27), cvpr (27), paper (27), vision (21), part (21), this (21), generation (21), 2023 (21), computer (19), objects (19), world (19), video (18), models (18), based (17), 2024 (17), are (16), high (16), learning (15), parts (15), 2026 (15), pattern (14), recognition (14), year (14), code (14), from (14), guibas (14), leonidas (14), title (13), author (13), bibtex (13), project (13), page (13), abstract (13), trained (13), generative (13), point (13), using (12), reconstruction (12), 2021 (12), accepted (12), inproceedings (11), shape (11), booktitle (11), scene (11), supervision (11), methods (11), neural (11), object (11), fidler (10), sanja (10), can (10), quality (10), data (10), wang (10), geiger (9), andreas (9), jun (9), shapes (9), which (9), demonstrate (9), more (9), single (9), primitives (9), autoregressive (9), aware (9), scenes (9), curvecloudnet (9), motion (9), workshop (9), representations (8), than (8), poster (8), latent (8), synthesis (8), park (8), jeong (8), joon (8), 2025 (8), may (8), proceedings (7), ieee (7), conf (7), complex (7), while (7), without (7), existing (7), work (7), foundation (7), physical (7), nvidia (7), omnidreams (7), research (7), feb (7), various (6), 2020 (6), these (6), novel (6), structure (6), slides (6), implicit (6), processing (6), diverse (6), applications (6), only (6), room (6), generates (6), alto (6), latents (6), international (6), conference (6), real (6), liu (6), diffusion (6), cosmos (6), arxiv (6), wei (6), time (6), outstanding (6), received (6), superquadrics (5), beyond (5), cuboids (5), 2019 (5), into (5), shapenet (5), well (5), image (5), level (5), progress (5), however (5), its (5), between (5), where (5), fidelity (5), primitive (5), both (5), any (5), post (5), evaluations (5), atiss (5), partial (5), training (5), transformer (5), grid (5), generate (5), content (5), such (5), conditioned (5), pre (5), open (5), policy (5), zhang (5), pasta (5), 3dv (5), two (5), motive (5), autonomous (5), driving (5), was (5), jan (5), ali (4), representation (4), expressive (4), fine (4), not (4), hierarchical (4), decomposition (4), set (4), geometry (4), their (4), datasets (4), capture (4), efficient (4), state (4), amlan (4), kar (4), transformers (4), indoor (4), systems (4), simulation (4), present (4), train (4), wetzstein (4), gordon (4), previous (4), one (4), school (4), each (4), over (4), partnerf (4), generating (4), editable (4), creation (4), cc3d (4), compositional (4), iccv (4), cloud (4), curve (4), like (4), priors (4), photorealistic (4), shen (4), platform (4), xindi (4), gao (4), lorraine (4), jonathan (4), icml (4), dynamic (4), closed (4), loop (4), papers (4), reviewer (4), award (4), organizing (4), revisited (3), parsing (3), month (3), has (3), solution (3), moreover (3), consistent (3), results (3), human (3), dataset (3), have (3), blog (3), unsupervised (3), luc (3), van (3), gool (3), distinct (3), networks (3), impressive (3), local (3), fewer (3), experiments (3), invertible (3), network (3), surface (3), neurips (3), ability (3), realistic (3), layouts (3), interactive (3), synthetic (3), environments (3), makes (3), layout (3), alternating (3), topologies (3), clouds (3), spatial (3), detail (3), thought (3), voxel (3), geometric (3), performance (3), several (3), require (3), appearance (3), improved (3), compared (3), andrea (3), introduce (3), synthesizes (3), view (3), most (3), modeling (3), controllable (3), visual (3), works (3), backbone (3), cad (3), via (3), adversarial (3), distillation (3), huang (3), jing (3), method (3), first (3), yuxuan (3), lin (3), chen (3), jay (3), zhangjie (3), fangyin (3), ren (3), huan (3), ling (3), laura (3), leal (3), seung (3), wook (3), kim (3), francesco (3), ferroni (3), yifan (3), next (3), attribution (3), temporal (3), influence (3), vehicle (3), mar (3), got (3), phd (3), 2022 (3), area (3), chair (3), wacv (3), ulusoy (2), osman (2), been (2), long (2), goal (2), problem (2), traditional (2), cuboid (2), being (2), loss (2), learns (2), parse (2), categories (2), poses (2), could (2), rgb (2), perceive (2), recent (2), when (2), input (2), focuses (2), considering (2), address (2), challenging (2), formulation (2), allows (2), recover (2), structural (2), form (2), simple (2), components (2), faust (2), facilitates (2), reasoning (2), abstractions (2), katharopoulos (2), angelos (2), led (2), geometries (2), semantically (2), arrangements (2), accurately (2), reconstruct (2), number (2), defines (2), inn (2), enables (2), computation (2), accurate (2), art (2), maria (2), shugrina (2), karsten (2), kreis (2), advances (2), information (2), unlocks (2), better (2), tools (2), architecture (2), plausible (2), contrast (2), prior (2), sequence (2), same (2), general (2), arrangement (2), permutation (2), conditioning (2), across (2), end (2), bounding (2), front (2), parameters (2), simpler (2), implement (2), zhen (2), zhou (2), shijie (2), you (2), suya (2), kadambi (2), achuta (2), identifies (2), another (2), decode (2), runtime (2), propose (2), lightweight (2), improvement (2), but (2), konstantinos (2), tertikas (2), boxiao (2), pan (2), mikaela (2), angelina (2), ioannis (2), emiris (2), yannis (2), avrithis (2), gave (2), rise (2), able (2), locally (2), control (2), produce (2), does (2), explicit (2), nerfs (2), operations (2), different (2), hard (2), ray (2), affect (2), others (2), bahmani (2), sherwin (2), yan (2), xingguang (2), tagliasacchi (2), images (2), gans (2), multiple (2), approach (2), new (2), process (2), stearns (2), colton (2), rempe (2), davis (2), alex (2), jiateng (2), mascha (2), sébastien (2), modern (2), sensors (2), backbones (2), rich (2), rely (2), size (2), sparse (2), large (2), alternatives (2), variety (2), past (2), wan (2), ziyu (2), ian (2), hongyu (2), bokui (2), xiang (2), xiaoyu (2), liao (2), increased (2), demand (2), mode (2), distribution (2), directly (2), diversity (2), pipeline (2), article (2), agarwal (2), niket (2), arslan (2), bala (2), maciej (2), balaji (2), yogesh (2), journal (2), preprint (2), needs (2), digital (2), twin (2), help (2), customized (2), curation (2), critical (2), zeng (2), david (2), taixé (2), songlin (2), cuboidal (2), blending (2), meshes (2), text (2), specific (2), torralba (2), antonio (2), russakovsky (2), olga (2), machine (2), role (2), centric (2), framework (2), use (2), tuning (2), clips (2), dynamics (2), scalable (2), improves (2), selected (2), oral (2), honorable (2), mention (2), basant (2), aarti (2), garcia (2), cobo (2), guillermo (2), turki (2), haithem (2), seo (2), jaewoo (2), lucas (2), james (2), jialiang (2), kai (2), tothova (2), katarina (2), xie (2), kevin (2), tyszkiewicz (2), lutio (2), riccardo (2), ruilong (2), tianchang (2), cao (2), tianshi (2), pfaff (2), tobias (2), lew (2), william (2), xuanchi (2), gojcic (2), zan (2), zian (2), policies (2), scenarios (2), environment (2), actions (2), simulator (2), sensor (2), simulators (2), highly (2), mid (2), autoregressively (2), action (2), videos (2), alpamayo (2), evaluating (2), show (2), also (2), serve (2), publications (2), 2018 (2), check (2), documentation (2), released (2), interning (2), summer (2), available (2), wonderful (2), dec (2), 1st (2), 2nd (2), vocabulary (2), understanding (2), eccv (2), wild (2), started (2), stanford (2), jul (2), very (2), senior (2), scientist (2), worked (2), university (2), intelligence (2), paschalidou2019cvpr, abstracting, parsimonious, standing, presents, goes, exploiting, atomic, elements, lead, parses, easier, learn, provide, analytical, chamfer, avoids, need, computational, expensive, reinforcement, iterative, prediction, superquadric, surreal, body, flexibility, capturing, details, modelled, paschalidou2020cvpr, humans, characterized, low, reflectance, connectivity, adjacency, symmetry, properties, convolutional, cnns, demonstrated, even, majority, recovering, relations, proposing, jointly, recovers, higher, binary, tree, represented, modeled, organization, indeed, about, paschalidou2021cvpr, extraction, parallel, seek, represent, due, simplicity, fail, small, trade, off, implements, homeomorphic, mappings, sphere, target, compute, inverse, mapping, homeomorphism, turn, function, mesh, additional, freihand, thus, simultaneously, achieve, geometrically, interpretable, reconstructions, order, magnitude, abstraction, podcast, paschalidou2021neurips, synthesize, furniture, automatically, many, creating, given, type, floor, plan, rooms, unordered, sets, argue, natural, generally, useful, fully, automatic, example, used, completion, specified, user, suggestions, enable, leverages, equivariance, invariant, orderings, labeled, boxes, four, types, consistently, addition, runs, faster, zhen2023cvpr, introduces, surfaces, noisy, encodings, important, encode, vector, triplane, tradeoffs, coarse, lose, frequency, preserve, difficult, suffer, sequentially, alternate, before, converging, easy, find, preserves, expressiveness, decoding, validate, recovery, observe, tertikas2023cvpr, edit, essential, property, unlock, achieved, cannot, textures, devise, defined, augmented, affine, transformation, editing, applying, transformations, mixing, etc, ensure, manipulable, enforce, assignment, rays, sure, color, determined, nerf, result, altering, approaches, relying, bahmani2023iccv, conditional, semantic, limit, applicability, aligned, focus, nature, devising, layoutbased, implementing, field, stronger, inductive, bias, created, gan, allowing, kitti, 360, comparison, stearns2024cvpr, depth, lidar, operate, sweeping, laser, beams, resulting, notable, structures, scheme, called, takes, advantage, inherent, discard, traversal, patterns, generic, parameterizes, collection, polylines, dubbed, establishing, ordering, points, along, curves, captures, efficiently, reason, evaluate, exhibit, outperforms, segmentation, settings, notably, scaling, exhibiting, all, handle, larger, wan2024cvpr, robotics, gaming, powerful, pipelines, capable, synthesizing, score, sampling, sds, algorithm, optimize, rendered, maintains, likelihood, evaluated, finding, correct, dimensional, produced, often, leads, issues, saturation, smoothing, janus, artifacts, paradigm, utilizes, instead, focusing, seeking, discrepancy, multi, renderings, manner, prompt, harnessing, space, wide, including, continuous, interpolation, domain, superiority, terms, agarwal2025arxiv, 2501, 03575, digitally, itself, developers, build, setups, position, purpose, tuned, downstream, covers, examples, tokenizers, builders, solve, problems, society, make, source, weight, permissive, licenses, artur, zolkowski, qingqing, zhao, qinsheng, xiaohui, yen, yang, jiashu, xinyue, ting, chun, heng, haoxiang, hao, jibin, varghese, cheng, tseng, przemek, tredak, lyne, tchapmi, shitao, tang, bartosz, stefaniak, stella, shi, schmerling, vasanth, rao, naik, sabavat, xiaowei, fitsum, reda, morteza, ramezanali, lindsey, pavao, zeeshan, patel, sriharsha, niverty, seungjun, nah, arsalan, mousavian, kaichun, hanzi, mao, qianli, alice, luo, xian, ming, tsung, hsuan, zhaoshuo, anqi, shiyi, lan, grace, lam, gergely, klár, jingyi, jin, pooya, jannaty, jacob, huffman, jiahui, ethan, siddharth, gururani, jinwei, yunhao, songwei, dieter, fox, michele, fenzi, jiaojiao, fan, daniel, dworakowski, ding, yin, cui, yongxin, prithvijit, chattopadhyay, tiffany, cai, erik, barker, li20263dv, automate, tremendous, deep, comprises, main, implemented, decoder, composes, sequences, stages, annotated, watertight, showcase, perform, inputs, scratch, guided, explicitly, box, boundaries, considers, underlying, select, meaningful, variations, evidenced, non, wu2026icml, taix, despite, rapid, influencing, poorly, understood, gradient, scales, study, improve, degrade, isolates, static, weighted, masks, yielding, strongly, guides, consistency, plausibility, smoothness, degree, vbench, achieving, preference, win, rate, pretrained, base, represents, attribute, rather, curate, basant2026arxiv, micha, 2606, 03159, capabilities, advance, safe, evaluation, tail, remains, bottleneck, actively, interacts, dynamically, update, generated, observations, offer, photorealism, they, fundamentally, constrained, initial, captured, struggle, generalize, overcome, limitations, leveraging, 21k, hours, unobserved, phenomena, extreme, weather, unpredictable, agent, behaviors, crucially, conditions, frames, current, immediate, deployed, system, alpasim, orchestrator, acts, responsive, reactive, providing, comprehensive, additionally, preliminary, indicating, wam, achieves, strong, vehicles, nurec, surpassing, vla, total, highlight, potential, architectures, michał, raynet, volumetric, potentials, pytorch, toronto, library, visualization, python, opengl, 3dviz, out, learned, cool, demo, london, fair, sep, after, years, successfully, defended, vancouver, second, tmlr, paris, seattle, milan, opensun3d, 3rd, challenges, ecommerce, april, postdoc, prof, copilot, honored, eecs, rising, stars, nov, three, multiphys, full, will, program, committee, eurographics, june, iclr, rematching, flow, presentation, news, previously, advised, postdoctoral, researcher, did, undergraduate, electrical, engineering, greece, during, lucky, spent, working, months, facebook, vedaldi, novotny, christos, diou, anastasios, delopoulos, aristotle, thessaloniki, max, planck, eth, center, santa, clara, graphics, building, spans, embodied, lab, presentations, home,
Text of the page (random words):
sical ai arxiv 2025 niket agarwal arslan ali maciej bala yogesh balaji erik barker tiffany cai prithvijit chattopadhyay yongxin chen yin cui yifan ding daniel dworakowski jiaojiao fan michele fenzi francesco ferroni sanja fidler dieter fox songwei ge yunhao ge jinwei gu siddharth gururani ethan he jiahui huang jacob huffman pooya jannaty jingyi jin seung wook kim gergely klár grace lam shiyi lan laura leal taixé anqi li zhaoshuo li chen hsuan lin tsung yi lin huan ling ming yu liu xian liu alice luo qianli ma hanzi mao kaichun mo arsalan mousavian seungjun nah sriharsha niverty david page despoina paschalidou zeeshan patel lindsey pavao morteza ramezanali fitsum reda xiaowei ren vasanth rao naik sabavat ed schmerling stella shi bartosz stefaniak shitao tang lyne tchapmi przemek tredak wei cheng tseng jibin varghese hao wang haoxiang wang heng wang ting chun wang fangyin wei xinyue wei jay zhangjie wu jiashu xu wei yang lin yen chen xiaohui zeng yu zeng jing zhang qinsheng zhang yuxuan zhang qingqing zhao artur zolkowski abstract project page paper code video bibtex physical ai needs to be trained digitally first it needs a digital twin of itself the policy model and a digital twin of the world the world model in this paper we present the cosmos world foundation model platform to help developers build customized world models for their physical ai setups we position a world foundation model as a general purpose world model that can be fine tuned into customized world models for downstream applications our platform covers a video curation pipeline pre trained world foundation models examples of post training of pre trained world foundation models and video tokenizers to help physical ai builders solve the most critical problems of our society we make cosmos open source and our models open weight with permissive licenses article agarwal2025arxiv title cosmos world foundation model platform for physical ai author nvidia and agarwal niket and ali arslan and bala maciej and balaji yogesh and others journal arxiv preprint arxiv 2501 03575 year 2025 cad photorealistic 3d generation via adversarial distillation computer vision and pattern recognition cvpr 2024 ziyu wan despoina paschalidou ian huang hongyu liu bokui shen xiaoyu xiang jing liao leonidas guibas abstract project page paper code video bibtex the increased demand for 3d data in ar vr robotics and gaming applications gave rise to powerful generative pipelines capable of synthesizing high quality 3d objects most of these models rely on the score distillation sampling sds algorithm to optimize a 3d representation such that the rendered image maintains a high likelihood as evaluated by a pre trained diffusion model however finding a correct mode in the high dimensional distribution produced by the diffusion model is challenging and often leads to issues such as over saturation over smoothing and janus like artifacts in this paper we propose a novel learning paradigm for 3d synthesis that utilizes pre trained diffusion models instead of focusing on mode seeking our method directly models the distribution discrepancy between multi view renderings and diffusion priors in an adversarial manner which unlocks the generation of high fidelity and photorealistic 3d content conditioned on a single image and prompt moreover by harnessing the latent space of gans and expressive diffusion model priors our method facilitates a wide variety of 3d applications including single view reconstruction high diversity generation and continuous 3d interpolation in the open domain the experiments demonstrate the superiority of our pipeline compared to previous works in terms of generation quality and diversity inproceedings wan2024cvpr title cad photorealistic 3d generation via adversarial distillation author wan ziyu and paschalidou despoina and huang ian and liu hongyu and shen bokui and xiang xiaoyu and liao jing and guibas leonidas booktitle proceedings ieee conf on computer vision and pattern recognition cvpr year 2024 curvecloudnet processing point clouds with 1d structure computer vision and pattern recognition cvpr 2024 colton stearns davis rempe alex fu jiateng liu sébastien mascha jeong joon park despoina paschalidou leonidas guibas abstract paper bibtex modern depth sensors such as lidar operate by sweeping laser beams across the scene resulting in a point cloud with notable 1d curve like structures in this work we introduce a new point cloud processing scheme and backbone called curvecloudnet which takes advantage of the curve like structure inherent to these sensors while existing backbones discard the rich 1d traversal patterns and rely on generic 3d operations curvecloudnet parameterizes the point cloud as a collection of polylines dubbed a curve cloud establishing a local surface aware ordering on the points by reasoning along curves curvecloudnet captures lightweight curve aware priors to efficiently and accurately reason in several diverse 3d environments we evaluate curvecloudnet on multiple synthetic and real datasets that exhibit distinct 3d size and structure we demonstrate that curvecloudnet outperforms both point based and sparse voxel backbones in various segmentation settings notably scaling to large scenes better than point based alternatives while exhibiting improved single object performance over sparse voxel alternatives in all curvecloudnet is an efficient and accurate backbone that can handle a larger variety of 3d environments than past works inproceedings stearns2024cvpr title curvecloudnet processing point clouds with 1d structure author stearns colton and rempe davis and fu alex and liu jiateng and mascha sébastien and park jeong joon and paschalidou despoina and guibas leonidas j booktitle proceedings ieee conf on computer vision and pattern recognition cvpr year 2024 cc3d layout conditioned generation of compositional 3d scenes international conference on computer vision iccv 2023 sherwin bahmani jeong joon park despoina paschalidou xingguang yan gordon wetzstein leonidas guibas andrea tagliasacchi abstract project page paper poster code bibtex in this work we introduce cc3d a conditional generative model that synthesizes complex 3d scenes conditioned on 2d semantic scene layouts trained using single view images different from most existing 3d gans that limit their applicability to aligned single objects we focus on generating complex scenes with multiple objects by modeling the compositional nature of 3d scenes by devising a 2d layoutbased approach for 3d synthesis and implementing a new 3d field representation with a stronger geometric inductive bias we have created a 3d gan that is both efficient and of high quality while allowing for a more controllable generation process our evaluations on synthetic 3d front and real world kitti 360 datasets demonstrate that our model generates scenes of improved visual and geometric quality in comparison to previous works inproceedings bahmani2023iccv author bahmani sherwin and park jeong joon and paschalidou despoina and yan xingguang and wetzstein gordon and guibas leonidas and tagliasacchi andrea title cc3d layout conditioned generation of compositional 3d scenes booktitle international conference on computer vision iccv year 2023 partnerf generating part aware editable 3d shapes without 3d supervision computer vision and pattern recognition cvpr 2023 konstantinos tertikas despoina paschalidou boxiao pan jeong joon park mikaela angelina uy ioannis emiris yannis avrithis leonidas guibas abstract project page paper poster slides code video bibtex impressive progress in generative models and implicit representations gave rise to methods that can generate 3d shapes of high quality however being able to locally control and edit shapes is another essential property that can unlock several content creation applications local control can be achieved with part aware models but existing methods require 3d supervision and cannot produce textures in this work we devise partnerf a novel part aware generative model for editable 3d shape synthesis that does not require any explicit 3d supervision our model generates objects as a set of locally defined nerfs augmented with an affine transformation this enables several editing operations such as applying transformations on parts mixing parts from different objects etc to ensure distinct manipulable parts we enforce a hard assignment of rays to parts that makes sure that the color of each ray is only determined by a single nerf as a result altering one part does not affect the appearance of the others evaluations on various shapenet categories demonstrate the ability of our model to generate editable 3d objects of improved fidelity compared to previous part based generative approaches that require 3d supervision or models relying on nerfs inproceedings tertikas2023cvpr author konstantinos tertikas and despoina paschalidou and boxiao pan and jeong joon park and mikaela angelina uy and ioannis emiris and yannis avrithis and leonidas guibas title partnerf generating part aware editable 3d shapes without 3d supervision booktitle proceedings ieee conf on computer vision and pattern recognition cvpr year 2023 alto alternating latent topologies for implicit 3d reconstruction computer vision and pattern recognition cvpr 2023 zhen wang shijie zhou jeong joon park despoina paschalidou suya you gordon wetzstein leonidas guibas achuta kadambi abstract project page paper poster slides code video bibtex this work introduces alternating latent topologies alto for high fidelity reconstruction of implicit 3d surfaces from noisy point clouds previous work identifies that the spatial arrangement of latent encodings is important to recover detail one school of thought is to encode a latent vector for each point point latents another school of thought is to project point latents into a grid grid latents which could be a voxel grid or triplane grid each school of thought has tradeoffs grid latents are coarse and lose high frequency detail in contrast point latents preserve detail however point latents are more difficult to decode into a surface and quality and runtime suffer in this paper we propose alto to sequentially alternate between geometric representations before converging to an easy to decode latent we find that this preserves spatial expressiveness and makes decoding lightweight we validate alto on implicit 3d recovery and observe not only a performance improvement over the state of the art but a runtime improvement of 3 10 inproceedings zhen2023cvpr title alto alternating latent topologies for implicit 3d reconstruction author wang zhen and zhou shijie and park jeong joon and paschalidou despoina and you suya and wetzstein gordon and guibas leonidas and kadambi achuta booktitle proceedings ieee conf on computer vision and pattern recognition cvpr year 2023 atiss autoregressive transformers for indoor scene synthesis advances in neural information processing systems neurips 2021 despoina paschalidou amlan kar maria shugrina karsten kreis andreas geiger sanja fidler abstract project page paper poster slides code video bibtex the ability to synthesize realistic and diverse indoor furniture layouts automatically or based on partial input unlocks many applications from better interactive 3d tools to data synthesis for training and simulation in this paper we present atiss a novel autoregressive transformer architecture for creating diverse and plausible synthetic indoor environments given only the room type and its floor plan in contrast to prior work which poses scene synthesis as sequence generation our model generates rooms as unordered sets of objects we argue that this formulation is more natural as it makes atiss generally useful beyond fully automatic room layout synthesis for example the same trained model can be used in interactive applications for general scene completion partial room re arrangement with any objects specified by the user as well as object suggestions for any partial room to enable this our model leverages the permutation equivariance of the transformer when conditioning on the partial scene and is trained to be permutation invariant across object orderings our model is trained end to end as an autoregressive generative model using only labeled 3d bounding boxes as supervision evaluations on four room types in the 3d front dataset demonstrate that our model consistently generates plausible room layouts that are more realistic than existing methods in addition it has fewer parameters is simpler to implement and train and runs up to 8x faster than existing methods inproceedings paschalidou2021neurips author despoina paschalidou and amlan kar and maria shugrina and karsten kreis and andreas geiger and sanja fidler title atiss autoregressive transformers for indoor scene synthesis booktitle advances in neural information processing systems neurips year 2021 neural parts learning expressive 3d shape abstractions with invertible neural networks computer vision and pattern recognition cvpr 2021 despoina paschalidou angelos katharopoulos andreas geiger sanja fidler abstract project page paper poster code blog slides video podcast bibtex impressive progress in 3d shape extraction led to representations that can capture object geometries with high fidelity in parallel primitive based methods seek to represent objects as semantically consistent part arrangements however due to the simplicity of existing primitive representations these methods fail to accurately reconstruct 3d shapes using a small number of primitives parts we address the trade off between reconstruction quality and number of parts with neural parts a novel 3d primitive representation that defines primitives using an invertible neural network inn which implements homeomorphic mappings between a sphere and the target object the inn allows us to compute the inverse mapping of the homeomorphism which in turn enables the efficient computation of both the implicit surface function of a primitive and its mesh without any additional post processing our model learns to parse 3d objects into semantically consistent part arrangements without any part level supervision evaluations on shapenet d faust and freihand demonstrate that our primitives can capture complex geometries and thus simultaneously achieve geometrically accurate as well as interpretable reconstructions using an order of magnitude fewer primitives than state of the art shape abstraction methods inproceedings paschalidou2021cvpr title neural parts learning expressive 3d shape abstractions with invertible neural networks author paschalidou despoina and katharopoulos angelos and geiger andreas and fidler sanja booktitle proceedings ieee conf on computer vision and pattern recognition cvpr month jun year 2021 learning unsupervised hierarchical part decomposition of 3d objects from a single rgb image computer vision and pattern recognition cvpr 2020 despoina pa...
|
