Abstract
Groupwise non-rigid registration is an important technique in medical image analysis. Recent studies show that its accuracy can be greatly improved by explicitly providing good initialisation. This is achieved by seeking a sparse correspondence using a parts+geometry model. In this paper we show that a single parts+geometry model is unlikely to establish consistent sparse correspondence for complex objects, and that better initialisation can be achieved using a set of models. We describe how to combine the strengths of multiple models, and demonstrate that the method gives state-of-the-art performance on three datasets, with the most significant improvement on the most challenging.
| Original language | English |
|---|---|
| Title of host publication | Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention |
| Pages | 156-163 |
| Number of pages | 8 |
| Volume | 15 |
| Edition | Pt 3 |
| Publication status | Published - 2012 Dec 1 |
Fingerprint
ASJC Scopus subject areas
- Medicine(all)
Cite this
Initialising groupwise non-rigid registration using multiple parts+geometry models. / Zhang, Pei; Yap, Pew Thian; Shen, Dinggang; Cootes, Timothy F.
Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention. Vol. 15 Pt 3. ed. 2012. p. 156-163.Research output: Chapter in Book/Report/Conference proceeding › Chapter
}
TY - CHAP
T1 - Initialising groupwise non-rigid registration using multiple parts+geometry models.
AU - Zhang, Pei
AU - Yap, Pew Thian
AU - Shen, Dinggang
AU - Cootes, Timothy F.
PY - 2012/12/1
Y1 - 2012/12/1
N2 - Groupwise non-rigid registration is an important technique in medical image analysis. Recent studies show that its accuracy can be greatly improved by explicitly providing good initialisation. This is achieved by seeking a sparse correspondence using a parts+geometry model. In this paper we show that a single parts+geometry model is unlikely to establish consistent sparse correspondence for complex objects, and that better initialisation can be achieved using a set of models. We describe how to combine the strengths of multiple models, and demonstrate that the method gives state-of-the-art performance on three datasets, with the most significant improvement on the most challenging.
AB - Groupwise non-rigid registration is an important technique in medical image analysis. Recent studies show that its accuracy can be greatly improved by explicitly providing good initialisation. This is achieved by seeking a sparse correspondence using a parts+geometry model. In this paper we show that a single parts+geometry model is unlikely to establish consistent sparse correspondence for complex objects, and that better initialisation can be achieved using a set of models. We describe how to combine the strengths of multiple models, and demonstrate that the method gives state-of-the-art performance on three datasets, with the most significant improvement on the most challenging.
UR - http://www.scopus.com/inward/record.url?scp=84872917202&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84872917202&partnerID=8YFLogxK
M3 - Chapter
C2 - 23286126
AN - SCOPUS:84872917202
VL - 15
SP - 156
EP - 163
BT - Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
ER -