Claudio Tortorici

Boosting 3D LBP-Based Face Recognition by Fusing Shape and Texture Descriptors on the Mesh

In this paper, we present a novel approach for fusing shape and texture local binary patterns (LBPs) on a mesh for 3D face recognition. Using a recently proposed framework, we compute LBP directly on the face mesh surface, then we construct a grid of the regions on the facial surface that can accommodate global and partial descriptions. Compared with its depth-image counterpart, our approach is distinguished by the following features: 1) inherits the intrinsic advantages of mesh surface (e.g., preservation of the full geometry); 2) does not require normalization; and 3) can accommodate partial matching. In addition, it allows early level fusion of texture and shape modalities. Through experiments conducted on the BU-3DFE and Bosphorus databases, we assess different variants of our approach with regard to facial expressions and missing data, also in comparison to the state-of-the-art solutions.

Representing 3D texture on mesh manifolds for retrieval and recognition applications

In this paper, we present and experiment a novel approach for representing texture of 3D mesh manifolds using local binary patterns (LBP). Using a recently proposed framework [37], we compute LBP directly on the mesh surface, either using geometric or photometric appearance. Compared to its depth-image counterpart, our approach is distinguished by the following features: a) inherits the intrinsic advantages of mesh surface (e.g., preservation of the full geometry); b) does not require normalization; c) can accommodate partial matching. In addition, it allows early-level fusion of the geometry and photometric texture modalities. Through experiments conducted on two application scenarios, namely, 3D texture retrieval and 3D face recognition, we assess the effectiveness of the proposed solution with respect to state of the art approaches.

Local binary patterns on triangular meshes

A novel framework for computing local binary patterns on triangular mesh manifolds.Mesh-LBP evidences uniformity and repeatability aspects.Mesh-LBP descriptors can cope with mesh irregularities and global deformations.Mesh-LBP can be deployed in both local and global shape analysis.Experiments on 3D texture classification, retrieval and 3D face recognition. In this paper, we introduce an original framework for computing local binary like-patterns on 2D mesh manifolds (i.e., surfaces in the 3D space). This framework, dubbed mesh-LBP, preservers the simplicity and the adaptability of the 2D LBP and has the capacity of handling both open and close mesh surfaces without requiring normalization as compared to its 2D counterpart. We describe the foundations and the construction of mesh-LBP and showcase the different LBP patterns that can be generated on the mesh. In the experimentation, we provide evidence of the uniform patterns in the mesh-LBP, the repeatability of its descriptors, and its robustness to moderate shape deformations. Then, we show how the mesh-LBP descriptors can be adapted to a number of surface local and global analysis including 3D texture classification and retrieval, and 3D face matching. We also compare the performance of the mesh-LBP descriptors with a bunch of state of the art surface descriptors.

Landmark detection from 3D mesh facial models for image-based analysis of dysmorphology

Facial landmark detection is a task of interest for facial dysmorphology, an important factor in the diagnosis of genetic conditions. In this paper, we propose a framework for feature points detection from 3D face images. The method is based on 3D Constrained Local Model (CLM) which learns both global variations in the 3D facial scan and local changes around every vertex landmark. Compared to state of the art methods our framework is distinguished by the following novel aspects: 1) It operates on facial surfaces, 2) It allows fusion of shape and color information on the mesh surface, 3) It introduces the use of LBP descriptors on the mesh. We showcase our landmarks detection framework on a set of scans including down syndrome and control cases. We also validate our method through a series of quantitative experiments conducted with the publicly available Bosphorus database.

Landmarks detection on 3D face scans using local histogram descriptors

In this work, we exploit 3D Constrained Local Model (CLM) for facial landmark detection. Our approach integrates the geometric information of 3D face scans. The fast increase demand of 3D data invite to develop 3D image processing methods for many applications and especially for automatic landmark detection. The new step in this paper is the introduction of mesh histogram of gradients (meshHOG) as local descriptors around every landmark location. The proposed work is evaluated on the publicly available Bosphorus database. A comparison with the other descriptors mesh LBP and mesh SIFT are also depicted.

Facial landmarks detection using 3D constrained local model on mesh manifold

This paper proposes a novel 3D Constrained Local Models (CLM) approach applied for the detection of facial landmarks in 3D images. This approach capitalizes on the properties of Independent Component Analysis (ICA) to define appropriate priors of a face Point Distribution Model. Tailored to the mesh manifold modality, this approach address the limitations of the depth images which require pose normalization and suffer from the loss of the shape information caused by 2D projection. We validate this framework through a series of experiments conducted with the public Bosporus database, whereby it demonstrates a competitive performance compared to other state of the art methods.

3D constrained local model with independent component analysis and non-Gaussian shape prior distribution: Application to 3D facial landmark detection

We present a novel statistical shape model and fitting process for the 3D Constrained Local Models (CLM), exploiting the properties of Independent Component Analysis (ICA), instead of the classic use of Principal Component Analysis (PCA), and adopting a non-Gaussian distribution of the shape prior information. Using ICA permits to exploit the real distribution of shape priors by adopting a Generalised Gaussian Distribution (GGD) model. Consequently, we derive a modified approach of the mean shift optimizer by using the Expectation-Maximization algorithms. We apply this novel method for the localization of face landmarks on 3D facial mesh models, which, to the best of our knowledge, is the first employment of the CLM variant on this kind of modality. Experiments conduced on the Bosphorus face database demonstrated that our approach outperforms state-of-the-art methods.

Early Features Fusion over 3D Face for Face Recognition

In this paper, a novel approach for fusing shape and texture Local Binary Patterns (LBP) for 3D Face Recognition is presented. Using the recently proposed mesh-LBP [23], it is now possible to compute LBP directly on a mesh manifold, allowing Early Feature Fusion to enhance face description power. Compared to its depth image counterparts, the proposed method (a) inherits the intrinsic advantages of mesh surfaces, (such as preservation of full geometry), (b) does not require face registration, (c) can accommodate partial or rotation matching, and (d) natively allows early-level fusion of texture and shape descriptors. The advantages of early-fusion is presented together with an experimentation of two merging schemes tested on the Bosphorus database.

Mesh LBP features for 3D constrained local model

We propose an automatic facial landmarks detection in 3D mesh manifold. The method is based on 3D Constrained Local Model (CLM) which learns both global variations in 3D face scan and local ones around every vertex landmark. Differently from the other approaches of CLM, our contribution is a full 3D mesh. The framework exploits the intrinsic 3D features around the mesh vertices by utilizing histogram-based mesh Local Binary Patterns (mesh-LBP). The experiments are conducted on publicly available 3D face scans Bosphorus database.

Boosting 3D LBP-based face recognition by fusing shape and texture descriptors on the mesh

In this paper, we present a novel approach for fusing shape and texture local binary patterns (LBP) for 3D face recognition. Using the framework proposed in [1], we compute LBP directly on the face mesh surface, then we construct a grid of the regions on the facial surface that can accommodate global and partial descriptions. Compared to its depth-image counterpart, our approach is distinguished by the following features: a) inherits the intrinsic advantages of mesh surface; b) does not require normalization; c) can accommodate partial matching. In addition, it allows early-level fusion of texture and shape modalities. Through experiments conducted on the BU-3DFE and Bosphorus databases, we assess different variants of our approach with regard to facial expressions and missing data.