#### DMCA

## Registration & Segmentation

### Citations

1046 | Energy Functions can be Minimized via Graph Cuts
- Kolmogorov, Zabih, et al.
- 2004
(Show Context)
Citation Context ... model consists of cliques of order two. The mapping of the other two terms is straightforward. It is possible to introduce high order cliques but such an energy cannot be (at least easily) minimized =-=[18]-=-, thus the global term will be approximated. We consider an approximation of the global cost that consists of assuming that for a given node p of a given deformation field/image i, the rest of the ima... |

695 | Nonrigid registration using free-form deformations: application to breast MR images
- Rueckert, LI, et al.
- 1999
(Show Context)
Citation Context ...mposing smoothness. The optimal parameters of the deformation should be determined by an energy minimization step. Gradient descent method is the most common approach to address the minimization task =-=[16]-=-, but suffers from important limitations. Among its limitations are the inability to guarantee the recovery of the global minimum and its computational inefficiency. Moreover, it is not modular since ... |

216 | Unbiased diffeomorphic atlas construction for computational anatomy. Neuroimage 2004;23:S151–60
- Joshi, Davis, et al.
(Show Context)
Citation Context ...tistics of the population. The idea of using the statistics of the population to construct a template based on which the similarity criterion is a natural extension of the above mentioned methods. In =-=[7]-=-, both the deformation and the template selection are optimized at the same time and a squared error dissimilarity measure was used. In [8] a normalized mutual information criterion in a Free Form Def... |

120 |
The Finite Element Method For Engineers
- Huebner, Thornton, et al.
- 1995
(Show Context)
Citation Context ... computations of the deformation field even more efficient and the same time localize the influence of the control points while at the same time being smooth enough, we considered the Hermite splines =-=[15]-=-. In this case equation 3 takes the following form: 1X 1X Di(xi) = (5) r=0 c=0 Hr(u)Hc(v)d p u 0 +r,v 0 +c i where Hl represents the lth basis function of the Hermite spline and the rest are defined a... |

117 | Learning from one example through shared densities on transformations
- Miller, Matsakis, et al.
- 2000
(Show Context)
Citation Context ...btained results. For example, the use of such methods becomes less evident when a important number of diseased patients are part of the population, or the population involves different modalities. In =-=[11]-=-, the idea of performing template-free group-wise registration was proposed where the objective criterion used is the sum of univariate entropies along pixel stacks. This method was applied in the pro... |

113 | Efficient graph-based energy minimization methods in computer vision
- Veksler
- 1999
(Show Context)
Citation Context ...ved through a compositional approach, where the final solution is obtained through successive optimization problems with respect to the deformation increment towards minimizing the objective function =-=[17, 1]-=-. Thus, by keeping the set of the labels in a reasonable size it becomes possible to approximate the optimal solution in an efficient way. (17) RR n° 683712 Sotiras & al. inria-00362960, version 3 - ... |

75 | Dense image registration through MRFs and efficient linear programming. Medical Image Analysis,
- Glocker, Komodakis, et al.
- 2008
(Show Context)
Citation Context ...ask in a great number of problems in computer vision and medical image processing. The goal is either to establish meaningful correspondences according to a certain criterion between a pair of images =-=[1]-=-, case of registration, or to convey information from a template, where prior information exists, to the matched image as in the case of the atlas-based segmentation problem [2]. Existing approaches t... |

48 | Consistent groupwise non-rigid registration for atlas construction,”
- Bhatia, Hajnal, et al.
- 2004
(Show Context)
Citation Context ...on is a natural extension of the above mentioned methods. In [7], both the deformation and the template selection are optimized at the same time and a squared error dissimilarity measure was used. In =-=[8]-=- a normalized mutual information criterion in a Free Form Deformation framework was used. The template was defined as the average transformation of the population. It should be noted that in approache... |

47 | Efficient population registration of 3D data
- Zollei, Learned-Miller, et al.
- 2005
(Show Context)
Citation Context ...ree group-wise registration was proposed where the objective criterion used is the sum of univariate entropies along pixel stacks. This method was applied in the problem of group-wise registration in =-=[12]-=- by combining affine transforms and a stochastic gradient descent optimization framework. This affine congealing framework was further extended by [13] to include Free Form Deformations. State of the ... |

45 | A unified informationtheoretic approach to groupwise non-rigid registration and model building.
- Twining, Cootes, et al.
- 2005
(Show Context)
Citation Context ...ation criterion in a Free Form Deformation framework was used. The template was defined as the average transformation of the population. It should be noted that in approaches as the ones presented in =-=[9, 10]-=- a minimum description length criterion was used to perform group-wise registration and model building. The main limitation of these methods is relevant to the ”use” of a template which could greatly ... |

43 |
Diffeomorphic registration using B-splines
- Rueckert, Aljabar, et al.
- 2006
(Show Context)
Citation Context ... constraint to be addressed refers to the diifeomorphic property of the proposed population registration framework. This can easily introduced by imposing hard constraints to the allowed deformations =-=[20]-=-. As the solution space is actually a quantized version of the deformation space and corresponds to the label set, it is simple to restrict the value that the maximum displacement can take in terms of... |

41 | Fast approximately optimal solutions for single and dynamic MRFs
- Komodakis, Tziritas
(Show Context)
Citation Context ...mize the successive MRFs, that is to assign a label l to all the nodes p of the constructed graph, a state of the art minimization method will be used. The method is called fast-PD and is detailed in =-=[19]-=-. The last constraint to be addressed refers to the diifeomorphic property of the proposed population registration framework. This can easily introduced by imposing hard constraints to the allowed def... |

38 | The role of image registration in brain mapping,”
- Toga, Thompson
- 2001
(Show Context)
Citation Context ...re made with the respect to it. Thus, in the case of atlas creation different reference images will result in different atlas, where each atlas is inherently biased towards the chosen reference image =-=[3]-=-. Such a behavior is the opposite to the one expected towards appropriate representation of the population. The previous remarks motivate an approach where all images should be let to deform mutually ... |

29 |
Groupwise non-rigid registration using polyharmonic clamped-plate splines,”
- Marsland, Twining, et al.
- 2003
(Show Context)
Citation Context ... work has been dedicated towards optimizing the selection of the reference image [4]. Once such a pose has been recovered, conventional registration methods based in pair-wise criteria can be applied =-=[5]-=-. The main drawback of this approach is that reference must be a member of the population. A natural way, to address this shortcoming is to perform simultaneously the registration of all subjects to a... |

29 |
Segmentation of Anatomical Structures in Chest Radiographs using Supervised Methods: A Comparative Study on a Public Database
- Ginneken, Stegmann, et al.
- 2006
(Show Context)
Citation Context ...nd 12 bit gray levels. 154 images contain exactly one pulmonary lung nodule each; the other 93 images contain no lung nodules. The images where segmented and made publicly available by the authors of =-=[23]-=-. From this publicly available data base only the 93 images that contain the no lung nodules were used. The images were scaled down to 256 × 256 and no further processing took place. From this data se... |

27 | Injectivity conditions of 2D and 3D uniform cubic B-spline functions,”
- Choi, Lee
- 2000
(Show Context)
Citation Context ...mum displacement can take in terms of the grid spacing. Towards imposing the diffeomorphic property on the deformation field, the maximum displacement is restricted to be 0 : 4 times the grid spacing =-=[21]-=- in the case of the cubic-B splines. Following [21], the bound for the maximum displacement, in the case of the cubic Hermite spline, is proven to be 0 : 25 times the grid spacing. 4 Experimental Vali... |

27 |
Development of a digital image database for chest radiographs with and without a lung nodule,‖
- Shiraishi, Katsuragawa, et al.
- 2000
(Show Context)
Citation Context ...t belong to patients as well as ones of healthy subjects are part of the data set 2. RR n° 683714 Sotiras & al. The third data set consists of chest radiographs that are taken from the JSRT database =-=[22]-=-. This database contains 247 PA chest radiographs collected from 13 institutions in Japan and one in the United States. The images were scanned from films to a size of 2048 × 2048 pixels, a spatial re... |

25 | Least biased target selection in probabilistic atlas construction
- Park, PH, et al.
(Show Context)
Citation Context ... than two. The major difference between all methods proposed lies in the definition of the reference frame. Significant work has been dedicated towards optimizing the selection of the reference image =-=[4]-=-. Once such a pose has been recovered, conventional registration methods based in pair-wise criteria can be applied [5]. The main drawback of this approach is that reference must be a member of the po... |

25 | Free-form B-spline deformation model for groupwise registration
- SK, Golland, et al.
- 2007
(Show Context)
Citation Context ...ied in the problem of group-wise registration in [12] by combining affine transforms and a stochastic gradient descent optimization framework. This affine congealing framework was further extended by =-=[13]-=- to include Free Form Deformations. State of the art population-registration methods have either opted for pair-wise registrations and template construction or global statistical measurements. Both co... |

19 |
Construction of a brain template from MR images using state-of-the-art registration and segmentation techniques”,
- Seghers, Agostino, et al.
- 2004
(Show Context)
Citation Context ... of all subjects to a common pose that is not predefined. INRIAMRF-based Diffeomorphic Population Deformable Registration & Segmentation 5 inria-00362960, version 3 - 8 Apr 2009 Towards this end, in =-=[6]-=- all the pair-wise registrations between members of the population were considered and a mean model was created by composing the mean deformations for each member into a mean deformation. This improve... |

9 | Targeted prostate biopsy using statistical image analysis - Zhan - 2007 |

7 | The impact of atlas formation methods on atlas-guided brain segmentation
- Zollei, Shenton, et al.
- 2007
(Show Context)
Citation Context ... to simultaneously segment all examples while imposing image-based consistencies between them. Such a concept is well motivated from a number of recent studies in the field of population registration =-=[24, 12, 8]-=- which demonstrate that pair-wise methods are quite biased and their performance is significant lower to the population ones. In this report, we introduce the concept of population segmentation throug... |

2 | Comparing pairwise and simultaneous joint registrations of decorrelating interval exams using entropic graphs
- Ma, Narayanan, et al.
(Show Context)
Citation Context ...omputational demand, memory requirements and running times in particular when addressing the deformable case. These constraints were partially addressed through efficient computational approximations =-=[12, 14]-=- but still when referring to the deformable case, the spectrum of these methods is far from being satisfactory. However, in terms of modularity with respect to the selected registration criterion most... |

1 |
Robust autonomus model learning from 2d and 3d data sets
- Langs, Donner, et al.
- 2007
(Show Context)
Citation Context ...ation criterion in a Free Form Deformation framework was used. The template was defined as the average transformation of the population. It should be noted that in approaches as the ones presented in =-=[9, 10]-=- a minimum description length criterion was used to perform group-wise registration and model building. The main limitation of these methods is relevant to the ”use” of a template which could greatly ... |