Meta tags:

Headings (most frequently used words):

motion, retargeting, in, 2d, and, learning, character, agnostic, for, abstract, decomposing, re, composing, skeleton, view, angle, interpolation, video, performance, cloning, retrival,

Text of the page (most frequently used words):
motion (21), and (18), the (16), retargeting (10), video (9), camera (8), from (7), view (7), our (5), human (5), #skeleton (5), angle (5), representation (4), different (4), which (4), videos (4), for (4), one (4), performance (4), with (4), extract (3), agnostic (3), latent (3), extracted (3), applied (3), skeletons (3), new (3), can (3), deep (3), three (3), are (3), captured (3), using (3), cloning (3), that (3), such (3), university (3), abstract (2), character (2), directly (2), may (2), then (2), other (2), possibly (2), train (2), neural (2), network (2), decompose (2), synthetic (2), into (2), similar (2), without (2), need (2), reconstruction (2), interpolation (2), axis (2), enables (2), novel (2), proportions (2), this (2), actor (2), search (2), dataset (2), wild (2), motions (2), applications (2), computer (2), existing (2), paper (2), poses (2), temporal (2), approach, ordinary, very, shown, viewpoints, well, projections, data, spaces, shuffled, composed, form, combinations, various, left, angles, right, horizontal, vertical, ability, perform, use, drive, done, recent, techniques, propose, generative, networks, produce, frames, contain, appearance, target, reenacting, driving, given, query, being, body, individual, retrival, decomposing, composing, analyzing, challenging, task, wide, variety, vision, graphics, application, particular, importance, animation, performer, another, while, humans, move, dimensions, vast, majority, requiring, pose, recovery, before, approaches, present, method, between, performers, explicitly, reconstruct, parameters, order, achieve, goal, learn, high, level, invariant, geometry, key, idea, sequences, components, having, able, combine, views, decode, retargeted, sequence, compare, ground, truth, demonstrate, framework, used, robustly, bypassing, outperforming, methods, when, also, additional, driven, cartoons, retrieval, code, corresponding, author, hebrew, peking, tel, aviv, beijing, film, academy, daniel, cohen, baoquan, chen, dani, lischinski, rundi, kfir, aberman, learning,

Text of the page (random words):
motion retargeting in 2d learning character agnostic motion for motion retargeting in 2d kfir aberman 1 2 rundi wu 3 dani lischinski 4 baoquan chen 3 daniel cohen or 2 1 beijing film academy 2 tel aviv university 3 peking university 4 hebrew university corresponding author abstract analyzing human motion is a challenging task with a wide variety of applications in computer vision and in graphics one such application of particular importance in computer animation is the retargeting of motion from one performer to another while humans move in three dimensions the vast majority of human motions are captured using video requiring 2d to 3d pose and camera recovery before existing retargeting approaches may be applied in this paper we present a new method for retargeting video captured motion between different human performers without the need to explicitly reconstruct 3d poses and or camera parameters in order to achieve our goal we learn to extract directly from a video a high level latent motion representation which is invariant to the skeleton geometry and the camera view our key idea is to train a deep neural network to decompose temporal sequences of 2d poses into three components motion skeleton and camera view angle having extracted such a representation we are able to re combine motion with novel skeletons and camera views and decode a retargeted temporal sequence which we compare to a ground truth from a synthetic dataset we demonstrate that our framework can be used to robustly extract human motion from videos bypassing 3d reconstruction and outperforming existing retargeting methods when applied to videos in the wild it also enables additional applications such as performance cloning video driven cartoons and motion retrieval paper video code motion retargeting in 2d our approach is to extract an abstract character and camera agnostic latent representation of human motion directly from ordinary video the extracted motion may then be applied to other possibly very different skeletons and or shown from new viewpoints which can be extracted as well from other videos decomposing and re composing we train a deep neural network to decompose 2d projections of synthetic 3d data into three latent spaces motion skeleton and camera view angle which are then shuffled and re composed to form new combinations skeleton and view angle retargeting retargeting of similar motion to various skeletons left and different view angles right without the need for 3d reconstruction interpolation interpolation of view angle horizontal axis and motion vertical axis video performance cloning the ability to perform motion retargeting in 2d enables one to use a video captured performance to drive a novel 2d skeleton with possibly different proportions this is done by using recent performance cloning techniques that propose a deep generative networks to produce frames that contain the appearance of a target actor reenacting the motion of a driving actor motion retrival using our motion representation we can search in a dataset of videos in the wild for motions similar to one in a video given as a query with the search being agnostic to the body proportions of the individual and the camera view angle