Majdi Jribi

A robust invariant bipolar representation for R 3 surfaces: applied to the face description

In this paper, we intend to introduce a novel invariant curved surface representation under the 3D motion group. It is constructed from the superposition of the two geodesic potentials generated from a given couple of surface points. By sampling this continuous representation, invariant points are extracted from a large neighborhood around these reference points. Different numerical methods are implemented in order to find an efficient approximation in the mean of the shape distance. The inference of small distortions of points positions applied to the reference points is analyzed. We apply the proposed representation to real 3D images. The experimentations are performed on the 3D facial database Bosphorus.

A Novel Geometrical Approach for a Rapid Estimation of the HARDI Signal in Diffusion MRI

In this paper, we address the problem of the diffusion signal reconstruction from a limited number of samples. The HARDI (High Angular Resolution Diffusion Imaging) technique was proposed as an alternative to resolve the problems of crossing fibers in the case of Diffusion Tensor Imaging (DTI). However, it requires a long scanning time for the acquisition of the Diffusion Weighted (DW) images. This fact makes hard the clinical applications. We propose here a novel geometrical approach to accurately estimate the HARDI signal from a few number of DW images. The missing diffusion data are obtained according to their neighborhood from a reduced set of diffusion directions on the sphere of the q-space. The experimentations are performed on both synthetic data and many digital phantoms simulating crossing fibers on the brain tissues. The obtained results show the accuracy of the reconstruction of the Fiber Orientation Distribution (FOD) function from the estimated diffusion signal.

3D Human Shapes Correspondence using the Principal Curvature Fields on a Local Surface Parametrization.

In this paper, we address the problem of the correspondence b etween 3D non-rigid human shapes. We propose a local surface description around the 3D human body extremi i s. It is based on the mean of principal curvature fields values on the intrinsic Darcyan parametriz ation constructed around these points. The similarity between the resulting descriptors is, then, measured in the sense of theL2 distance. Experiments on a several human objects from the TOSCA dataset confirm the accuracy of t he proposed approach.

A Novel and Accurate Local 3D Representation for Face Recognition

In this paper, we intend to introduce a novel curved 3D face representation. It is constructed on some static parts of the face which correspond to the nose and the eyes. Each part is described by the level curves of the superposition of several geodesic potentials generated from many reference points. We propose to describe the eye region by a bipolar representation based on the superposition of two geodesic potentials generated from two reference points and the nose by a three-polar one (three reference points). We use the BU-3DFE database of 3D faces to test the accuracy of the proposed approach. The obtained results in the sense of the Hausdorff shape distance prove the performance of the novel representation for 3D faces identification. The obtained scores are comparable to the state of the art methods in the most of cases.

Accurate 3D Shape Correspondence by a Local Description Darcyan Principal Curvature Fields

In this paper, we propose a novel approach for finding correspondence between three-dimensional shapes undergoing non-rigid transformations. Our proposal is based on the computation of the mean of curvature fields values on a local parametrization constructed around interest points on the surface. This local parametrization corresponds to the Darcyan coordinates system. Thereafter, correspondence is found by measuring the \(L_{2}\) distance between obtained descriptors. We conduct the experimentation on the full objects of the Tosca database which contains a set of 3D objects with non-rigid deformations. The obtained results show the performance of the proposed approach.

Enhancing 3D Face Recognition by a Robust Version of ICP Based on the Three Polar Representation

In this paper, we intend to propose a framework for the description and the matching of three dimensional faces. Our starting point is the representation of the 3D face by an invariant description under the M(3) group of translations and rotations. This representation is materialized by the points of the arc-length reparametrization of all the level curves of the three polar representation. These points are indexed by their level curve number and their position in each level. With this type of description we need a step of registration to align 3D faces with different expressions. Therefore, we propose to use a robust version of the iterative closest point algorithm (ICP) adopted to 3D face recognition context. We test the accuracy of our approach on a part of the BU-3DFE database of 3D faces. The obtained results for many protocols of the identification scenario show the performance of such framework.

A Novel Canonical Form for the Registration of Non Rigid 3D Shapes

In this paper, we address the problem of non rigid 3D shapes registration. We propose to construct a canonical form for the 3D objects corresponding to the same shape with different non rigid inelastic deformations. It consists on replacing the geodesic distances computed from three reference points of the original surface by the Euclidean ones calculated from three points of the novel canonical form. Therefore, the problem of non rigid registration is transformed to a rigid matching between canonical forms. The effectiveness of such method for the recognition and the retrieval processes is evaluated by the experimentation on the TOSCA database objects in the mean of the Hausdorff Shape distance.

A robust and isotropic curved surface representation for 3D faces description

A novel curved surface representation which is relatively invariant under the transformations of the 3D motion group is introduced in this paper. It is computed from the superposition of the two geodesic potentials generated from a given couple of surface points. By considering a levels set of such geodesic potentials, finite invariant points are obtained. Two numerical methods are implemented and compared in order to find an efficient approximation of this representation in the sense of the Hausdorff shape distance. Its robustness under an imprecision of reference points positions is studied. Experimentations are performed on the 3D real faces of the database Bosphorus.

A New Bipolar Surfaces Representation: Approximation on 3D Meshes

In this present work, we propose two approximation methods of a novel curved surface representation which is relatively invariant under 3D motion group, introduced in our recent work [11]. It is constructed from the superposition of two geodesic potentials of two fixed surface points. From this representation, finite invariant points could be extracted locally in the neighborhood of such two reference points. By comparing each approximation method with a reference representation computed on an analytical surface, the best method is retained by the way of the Hausdorff distance in the shape space. Subsequently, this approximation algorithm is applied to analyze the shape of 3D human faces. Its discrimination power has been identified by experimentations performed on the database Bosphorus.

Unsupervised classifier based on geodesic invariant 3D curve for face surfaces analysis

Here, we intend to introduce new face invariant descriptors, composed by two kinds of features, in order to explore the problem of faces classification. The first kind is defined from the p-order moments of a curvature function of the geodesic curve according to its arc length. The second one describes relative positions between important localities of faces. Two classes Fisher discriminate analysis is applied for a dimension reduction. A two dimensional multi classes Expectation Maximization algorithm (2D-EM) is used to identify the components of the mixture distribution. Then, the classification is obtained after applying the Bayes decision rule which is the most optimal for the minimization of the classification error. Such classification gives the sub groups having homogenous similar faces.