Toward Pose-Invariant 2-D Face Recognition Through Point Distribution Models and Facial Symmetry

Abstract

This paper proposes novel ways to deal with pose variations in a 2-D face recognition scenario. Using a training set of sparse face meshes, we built a Point Distribution Model and identified the parameters which are responsible for controlling the apparent changes in shape due to turning and nodding the head, namely the pose parameters. Based on them, we propose two approaches for pose correction: 1) a method in which the pose parameters from both meshes are set to typical values of frontal faces, and 2) a method in which one mesh adopts the pose parameters of the other one. Finally, we obtain pose corrected meshes and, taking advantage of facial symmetry, virtual views are synthesized via Thin Plate Splines-based warping. Given that the corrected images are not embedded into a constant reference frame, holistic methods are not suitable for feature extraction. Instead, the virtual faces are fed into a system that makes use of Gabor filtering for recognition. Unlike other approaches that warp faces onto a mean shape, we show that if only pose parameters are modified, client specific information remains in the warped image and discrimination between subjects is more reliable. Statistical analysis of the authentication results obtained on the XM2VTS database confirm the hypothesis. Also, the CMU PIE database is used to assess the performance of the proposed methods in an identification scenario where large pose variations are present, achieving state-of-the-art results and outperforming both research and commercial techniques.Index Terms-CMU PIE database, facial symmetry, Gabor jets, point distribution models, pose-invariant face recognition, thin-plate splines, XM2VTS database.

Hierarchical Image Matching for Pose-invariant Face Recognition

“…In Table 2 the results of a comparison of the proposed approach with two other methods are presented. It is evident from the table, the proposed approach outperforms the method in [5] in all poses except pose 05. The overall performance of the proposed method is far better than the one in [5].…”

“…It is evident from the table, the proposed approach outperforms the method in [5] in all poses except pose 05. The overall performance of the proposed method is far better than the one in [5]. Also, our method outperforms the state-of-the-art method in [22] in four different poses, especially when there is a large deviation in viewing angle from frontal.…”

“…The frontal views of subjects (pose 27) are considered as gallery images and all the rest (12 different poses) are used as test images. Table 2: Comparison of correct recognition percentage across different poses between the proposed method and the works in [5] and [22]. Figure 5 shows the result of recognition using texture and also shape and texture together.…”

“…One of the most well known approaches in 2D is the active appearance model by Cootes et al [4]. Another example is the work by Jimenez and Castro [5] in which point distribution models are constructed and eigenvectors responsible for appearance changes due to change of head pose are identified. After correcting the pose, Gabor filtering is used for recognition.…”

Multi-attribute robust facial feature localization

“…These are called primary in that they are characterized by corners and edges, and they are most instrumental in determining facial identity and expression. There are also secondary landmark points, such as nostrils, chin, nose bridge, cheek contours as many 62 points as in [11]. These are called secondary in that they have more scarce low-level image evidence.…”