Light-invariant fitting of active appearance models

Pizarro, Daniel; Peyras, Julien; Bartoli, Adrien

doi:10.1109/cvpr.2008.4587651

Cited by 15 publications

(8 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using an explicit local illumination model would probably cause instabilities: Local intensity variations would both be explained by the warp and the illumination model. A solution might be to use invariant images as proposed in [25] and very recently to Active Appearance Models in [26]. The efficiency of the proposed approach could also be improved by using the ESM algorithm for the minimization.…”

Section: Resultsmentioning

confidence: 96%

Direct Estimation of Nonrigid Registrations with Image-Based Self-Occlusion Reasoning

Gripon

Bartoli

Sayd

2010

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

The registration problem for images of a deforming surface has been well studied. External occlusions are usually well handled. In 2D image-based registration, self-occlusions are more challenging. Consequently, the surface is usually assumed to be only slightly self-occluding. This paper is about image-based nonrigid registration with self-occlusion reasoning. A specific framework explicitly modeling self-occlusions is proposed. It is combined with an intensity-based, "direct" data term for registration. Self-occlusions are detected as shrinkage areas in the 2D warp. Experimental results on several challenging data sets show that our approach successfully registers images with self-occlusions while effectively detecting the self-occluded regions.

show abstract

Section: Resultsmentioning

confidence: 96%

Direct Estimation of Nonrigid Registrations with Image-Based Self-Occlusion Reasoning

Gripon

Bartoli

Sayd

2010

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

show abstract

“…Examples include template matching handled in the frequency domain [6], modifications to correlation-based statistics [13,16], conquer-and-divide approaches to pixel space by breaking down the entire pixel space into smaller subspaces and subjecting them to localised matching [17], or even methods of matching stereo images where richer problem spaces are considered [2]. More recent alternative global or subglobal methods [11,12] rely on a number of restrictive assumptions and thus cannot handle more general non-homogeneous deviations. At present, to the best of our knowledge, no matching score for loosely constrained non-homogeneous contrast and offset deviations has been proposed 1 .…”

Section: Previous Workmentioning

confidence: 99%

Image Matching with Spatially Variant Contrast and Offset: A Quadratic Programming Approach

Shorin

Gimel’farb

Delmas

et al. 2008

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Template matching is widely used in machine vision, digital photogrammetry, and multimedia data mining to search for a target object by similarity between its prototype image (template) and a sensed image of a natural scene containing the target. In the real-world environment, similarity scores are frequently affected by contrast / offset deviations between the template and target signals. Most of the popular least-squares scores presume only simple smooth deviations that can be approximated with a low-order polynomial. This paper proposes an alternative and more general quadratic programming based matching score that extends the conventional least-squares framework onto both smooth and non-smooth signal deviations.

show abstract

“…This model fails in the presence of illumination changes and several solutions have been proposed to address this issue such as structure-texture decomposition [17], LightInvariant [11] or CENSUS [12,18] transforms. Our feature-based data term is independent of the direct term used so we restrict ourselves to the simple model (14) for clarity.…”

Section: Priorsmentioning

confidence: 99%

Combining Features and Intensity for Wide-Baseline Non-Rigid Surface Registration

Braux-Zin¹,

Dupont²,

Bartoli³

2013

Procedings of the British Machine Vision Conference 2013

Self Cite

View full text Add to dashboard Cite

Non-rigid surface registration consists in estimating the deformation of a known flat surface, usually by fitting a warp such as a Thin-Plate Spline or a Free-Form Deformation. Common techniques are split in two categories: (1) pixel-based surface tracking where important deformations can be estimated over a video sequence as long as the frame to frame steps are small and (2) wide-baseline feature-based registration, able to directly handle large deformations at the cost of reduced accuracy. We introduce a new direct data term robustly merging feature and pixel-based costs in a pyramidal variational approach. By using a robust estimator we achieve an implicit optimal filtering of features and automatic balancing between the two terms.

show abstract

Light-invariant fitting of active appearance models

Cited by 15 publications

References 9 publications

Direct Estimation of Nonrigid Registrations with Image-Based Self-Occlusion Reasoning

Direct Estimation of Nonrigid Registrations with Image-Based Self-Occlusion Reasoning

Image Matching with Spatially Variant Contrast and Offset: A Quadratic Programming Approach

Combining Features and Intensity for Wide-Baseline Non-Rigid Surface Registration

Contact Info

Product

Resources

About