Abstract
For man, several types of signals are available to recognize kinship, such as facial resemblance between families. Some studies show that we are able to recognize the kinship between a child's face and a face of a parent. The purpose of our work is to study the face-kin verification by using the Statistical Local Binary Patterns (Stat LBP) descriptor for the extraction of the local characteristics of facial images and the SIEDA (Side-Information based Exponential Discriminant Analysis) for the dimensionality reduction and classification of data. The extensive experimental evaluation carried out on Cornell kinshep database indicates that the proposed approach performs significantly better than state-of-the-art approaches.
Citation
Ouamane A. Stat LBP Feature Extraction and SIEDA Dimensionality Reduction and Classification for Face-Kin Verification. SM J Biometrics Biostat. 2017; 2(2): 1010