Enhanced Kinship Verification with Joint Color-Texture Features

Facial kinship verification, the process that determines whether or not a family relationship exists based on the analysis of facial images, is a subject of growing interest due to the diversity and the importance of its practical applications. This work explores an approach to kinship verification that uses a feature extraction technique called the Histogram of Local Binary Patterns (Hist LBP). The Hist LBP provides a richer representation of facial features by capturing both texture and color information. This study also investigates the effect of a power transform normalization technique, applied to the extracted features. The normalized Hist LBP features are then fed into a Support Vector Machine (SVM) classifier, trained using 5-fold cross-validation for robust model training and evaluation. The system’s performance was evaluated on the KinFaceW-II benchmark, achieving an accuracy of 91.15% with the optimal parameter configuration. This result is considered promising compared to many state-of-the-art approaches.