Multi-View Graph Fusion for Semi-Supervised Learning: Application to Image-Based Face Beauty Prediction

Facial Beauty Prediction (FBP) is an important visual recognition problem to evaluate the attractiveness of faces according human perception. Most existing FBP methods are based on supervised solutions using geometric or deep features. Semi-supervised learning for almost unexplored research area. In this work, we propose a graph-based semi-supervised method in which multiple graphs constructed find appropriate graph representation face images (with and without scores). The proposed combines both feature-based produce high-level instead single descriptor also improves discriminative ability score propagation methods. addition data graph, our approach fuses additional adaptively built predicted beauty values. Experimental results SCUTFBP-5500 facial dataset demonstrate superiority algorithm compared other state-of-the-art