Expression cloning

Noh, Junyong; Neumann, Ulrich

doi:10.1145/383259.383290

Cited by 226 publications

(88 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This proposed method is fully automatic, except for the initial fitting step on the first frame. (Automatic initial fitting can be achieved using the automated correspondence selection technique (Noh and Neumann 2001) but it is outside the scope of this paper.) We first describe the global alignment of 3D scans, followed by a description of the registration algorithm based on harmonic mapping and an iterative refinement scheme using local features.…”

Section: The Non-rigid Tracking Algorithmmentioning

confidence: 99%

“…c The generic face model with manually selected feature points (marked as red dots). (Automatic initial fitting can be achieved using the automated correspondence selection technique (Noh and Neumann 2001) but it is outside the scope of this paper.) d The result of the initial fitting to a 3D face scan data However, since the dense data samples in these 3D face scans are not registered in object space, inter-frame correspondences can not be established.…”

Section: Data Preparation and Initializationmentioning

confidence: 99%

“…(Witkin et al 1987;Black and Yacoob 1995;Essa and Pentland 1997;Lien et al 1998;Akgul and Kambhamettu 1999;Brand and Bhotika 2001;Torresani et al 2001;Gokturk et al 2001;Rittscher et al 2002;Goldenstein et al 2003;Wen and Huang 2003;Tomasi et al 2003;Ramanan and Forsyth 2003). Previous methods establishing 3D interframe correspondences for non-rigid motion largely fall into two categories: One depends on markers attached to the object (Guenter et al 1998;Kalberer and Van Gool 2001;Allen et al 2003) or on feature correspondences manually selected by the users (Noh and Neumann 2001); the other calculates correspondences based on the geometry using a 3D deformable/morphable model (Essa and Pentland 1997;Basu et al 1998;Pighin et al 1999;Tao and Huang 1999;DeCarlo and Metaxas 2002;Goldenstein et al 2003;Yezzi and Soatto 2003;Chai et al 2003;Blanz and Vetter 2003;Xiao et al 2004;Dimitrijevic et al 2004), or other 3D shape registration algorithms such as (Chen and Medioni 1991;Besl and McKay 1992;Zhang 1994). In general, most of the existing methods rely on templates with relatively few degrees of freedom.…”

Section: Introduction and Previous Workmentioning

confidence: 99%

“…This proposed method is fully automatic, except for the initial fitting step on the first frame. (Automatic initial fitting can be achieved using the automated correspondence selection technique (Noh and Neumann 2001) but it is outside the scope of this paper.) High quality dense point clouds of facial geometry moving at video speeds are acquired using a phase-shifting structured light ranging technique .…”

Section: Introduction and Previous Workmentioning

confidence: 99%

See 3 more Smart Citations

High Resolution Tracking of Non-Rigid Motion of Densely Sampled 3D Data Using Harmonic Maps

et al. 2007

View full text Add to dashboard Cite

We present a novel automatic method for high resolution, non-rigid dense 3D point tracking. High quality dense point clouds of non-rigid geometry moving at video speeds are acquired using a phase-shifting structured light ranging technique. To use such data for the temporal study of subtle motions such as those seen in facial expressions, an efficient non-rigid 3D motion tracking algorithm is needed to establish inter-frame correspondences. The novelty of this paper is the development of an algorithmic framework for 3D tracking that unifies tracking of intensity and geometric features, using harmonic maps with added feature correspondence constraints. While the previous uses of harmonic maps provided only global alignment, the proposed introduction of interior feature constraints allows to track non-rigid deformations accurately as well. The harmonic map between two topological disks is a diffeomorphism with minimal stretching energy and bounded angle distortion. The map is stable, insensitive to resolution changes and is robust to noise. Due to the strong implicit and explicit smoothness constraints imposed by the algorithm and the high-resolution data, the resulting registration/deformation field is smooth, continuous and gives dense one-to-one interframe correspondences. Our method is validated through a series of experiments demonstrating its accuracy and efficiency.

show abstract

Section: The Non-rigid Tracking Algorithmmentioning

confidence: 99%

Section: Data Preparation and Initializationmentioning

confidence: 99%

Section: Introduction and Previous Workmentioning

confidence: 99%

“…This proposed method is fully automatic, except for the initial fitting step on the first frame. (Automatic initial fitting can be achieved using the automated correspondence selection technique (Noh and Neumann 2001) but it is outside the scope of this paper.) High quality dense point clouds of facial geometry moving at video speeds are acquired using a phase-shifting structured light ranging technique .…”

Section: Introduction and Previous Workmentioning

confidence: 99%

See 2 more Smart Citations

High Resolution Tracking of Non-Rigid Motion of Densely Sampled 3D Data Using Harmonic Maps

et al. 2007

View full text Add to dashboard Cite

show abstract

“…Most works assume either (some of) the correspondences are known, or that the objects are already roughly aligned. For example, a few correspondences can be provided a priori, obtained either manually [26], [27] or by attaching markers to the object [4], [15], [19]. There is also a lot of research on the articulated motion, such as tracking human bodies or hand gestures [18], [43], which is similar to the type of semi-rigid deformations we consider.…”

mentioning

confidence: 99%

Toward Unsupervised Segmentation of Semi-Rigid Low-Resolution Molecular Surfaces

Guibas

Wang

2007

Algorithmica

View full text Add to dashboard Cite

In this paper we study a particular type of surface segmentation problem motivated by molecular biology applications. In particular, two input surfaces are given, coarsely modeling two different conformations of a molecule undergoing a semi-rigid deformation. The molecule consists of two subunits that move in a roughly rigid manner. The goal is to segment the input surfaces into these semi-rigid subcomponents. The problem is closely related to non-rigid surface registration problems, although considering only a special type of deformation that exists commonly in macromolecular movements (such as the popular hinge motion). We present and implement an efficient paradigm for this problem, which combines several existing and new ideas. We demonstrate the performance of our new algorithm by some preliminary experimental results in segmenting low-resolution molecular surfaces.

show abstract