Affine Alignment of Occluded Shapes

Sánta, Zsolt; Kató, Zoltán

doi:10.1109/icpr.2014.375

Cited by 2 publications

(16 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The work has addressed two different problems related to the algebraic framework introduced in the previous chapter. In the first problem, the method's tolerance against larger segmentation errors has been investigated [Sánta and Kato, 2014].…”

Section: Chapter 4 Robust Registration Of 2d Imagesmentioning

confidence: 99%

“…In a less controlled environment (e.g. processing images of surveillance systems), however, we always have to deal with different types of noise and occlusions [Sánta and Kato, 2014].…”

Section: Chapter Robust Registration Of 2d Imagesmentioning

confidence: 99%

“…In order to adapt the proposed framework above, let us assume that pursued transformation is defined by a homogeneous affine matrix A ∈ R 3×3 : A full affine transformation can be expressed using a single matrix multiplication between the homogeneous coordinates of the template and observation denoted byx recovery occlusion physical deformation While this relation is true when the observation is a perfect image of the template, in the presence of occlusions this will not be valid anymore. Let us denote the occluded regions of the observation by F occ (see Figure 4.1 for an example) [Sánta and Kato, 2014]. Then, we can define an object level identity relation as:…”

Section: Affine Alignment Of Occluded Shapesmentioning

confidence: 99%

“…, ℓ. For the sake of simplicity, we will assume that the ω i functions are acting on the Cartesian coordinates of the points [Sánta and Kato, 2014]. Observe that for all non-zero A affine transformations, the following equation is valid:…”

Section: Affine Alignment Of Occluded Shapesmentioning

confidence: 99%

“…, ℓ. Similarly to the affine framework proposed in [122], we use simple power functions for the {ω i } ℓ i=1 set and choosing ℓ ≥ 6 yields an overdetermined system, which is then solved in the least-squares sense [Sánta and Kato, 2014].…”

Section: Affine Alignment Of Occluded Shapesmentioning

confidence: 99%

See 4 more Smart Citations

Non-Rigid Registration of Visual Objects

Sánta

View full text Add to dashboard Cite

Sample results on the Bosphorus dataset. In each row, we show the template, the observation, the warped template and the landmark locations, respectively. In the last column, green denotes the ground truth position and red is the estimated location. While the proposed approach achieved good results near nose and mouth areas, the highest errors are near the eyebrows.. .. .. .. . 53 LIST OF FIGURES v 3.20 Comparison between the proposed, the GMMREG [5] and the CPD [6] algorithms. On the left diagram, the D GT measure is presented, while on the right we show the running times of each algorithm. The GMMREG slightly outperformed the proposed approach and the CPD algorithm in terms of alignment accuracy, but the proposed approach has the best running time.. .. .. .. .

show abstract

Section: Chapter 4 Robust Registration Of 2d Imagesmentioning

confidence: 99%

“…In a less controlled environment (e.g. processing images of surveillance systems), however, we always have to deal with different types of noise and occlusions [Sánta and Kato, 2014].…”

Section: Chapter Robust Registration Of 2d Imagesmentioning

confidence: 99%

Section: Affine Alignment Of Occluded Shapesmentioning

confidence: 99%

Section: Affine Alignment Of Occluded Shapesmentioning

confidence: 99%

Section: Affine Alignment Of Occluded Shapesmentioning

confidence: 99%

See 3 more Smart Citations