A locally constrained radial basis function for registration and warping of images

Siddiqui, Adil Masood; Masood, Asif; Saleem, Muhammad Qaiser

doi:10.1016/j.patrec.2008.11.004

Cited by 23 publications

(25 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For smooth, piecewise monotonic, compact supported functions (this is the most interesting case, as stated in (Wendland, 1995;Fornefett et al, 2001;Siddiqui et al, 2009)), the first equality in the previous equation holds not only in the sense of the quadratic norm, but also punctually for each x p À x q . It is very easy to notice that the integrand in the previous equation is nonnegative: e…”

Section: Proof Of Propositionmentioning

confidence: 99%

“…Since we cannot guarantee that K is PD for all possible scenarios, we cannot assure either that its determinant is non-zero, so it is possible that K is not invertible in some situations, and the interpolation problem has not a unique solution. Siddiqui et al (2009) It is stated in (Siddiqui et al, 2009) that to prove that /(r) is positive definite, we use Bochner's theorem. According to the previous discussion, Bochner's theorem is not necessary to show that a function is PD.…”

Section: Variable-width Rbf Kernelsmentioning

confidence: 99%

“…The nicest property of the function, however, is its equal distribution inside its support. This feature can be described as the small deviation of the slope of the function along its support, and translates in an equal amount of impact on interpolating pixel points on both sides of target landmark Siddiqui et al (2009). It is said in (Siddiqui et al, 2009) that the function in Eq.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Comments on: A locally constrained radial basis function for registration and warping of images

Tristán‐Vega

García-Pérez

2011

Pattern Recognition Letters

View full text Add to dashboard Cite

Section: Proof Of Propositionmentioning

confidence: 99%

Section: Variable-width Rbf Kernelsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Comments on: A locally constrained radial basis function for registration and warping of images

Tristán‐Vega

García-Pérez

2011

Pattern Recognition Letters

View full text Add to dashboard Cite

“…This allows us to obtain a mapping between each pixel, and so, to know the difference of illumination at each point of the face surface. The deformed image I 0 d is obtained using a Radial Basis Functions (RBF) scattered data interpolation [23] between two sets of landmarks manually picked on the two images I and I 0 . The ratio image R is computed by dividing I to I 0 d :…”

Section: Normal Map Generationmentioning

confidence: 99%

Easy acquisition and real‐time animation of facial wrinkles

Dutreve

Meyer

Bouakaz

2011

Computer Animation & Virtual

View full text Add to dashboard Cite

Facial animation details like wrinkles or bulges are very useful for the analysis and the interpretation of facial emotions and expressions. However, outfitting a virtual face with expression details for real-time applications is a difficult task. In this paper, we propose a mono-camera acquisition technique of facial animation details and a technique which add a wrinkle map layer (fine-scale animation) to a skinning layer (large-scale animation) for real-time rendering of a virtual 3D face. The acquisition is based on ratio image computed from two pictures of a same face, with and without expression. The real-time dynamic wrinkles technique is based on a small set of reference poses. These two methods offer an easy and lowcost way to capture facial animation details and use it for real-time facial animation.

show abstract

“…Examples include tissue edge distances as measured by the Haussdorff distance, 11 comparing the resulting transformations with the criterion standard reference transformations, 12 using the amount of tissue overlap for equivalent regions, 6,13 using imagesimilarity measures calculated between images, 14 and using visual assessments by human observers. 15 Extensive neuroimaging algorithm comparison studies have shown that registration performance is minimally affected by the many variations of labeling protocols and overlap measures.…”

mentioning

confidence: 99%