Global face reconstruction for face hallucination using orthogonal canonical correlation analysis

In this paper, a global face reconstruction framework for face hallucination is proposed to globally reconstruct a high-resolution (HR) version of a face from an input low-resolution (LR) face, based on learning from LR-HR face pairs using orthogonal canonical correlation analysis (orthogonal CCA). In our proposed algorithm, face images are first represented using principal component analysis (PCA). CCA with the orthogonality property is then employed to maximize the correlation between the PCA coefficients of the LR and the HR face pairs so as to improve the hallucination performance. The original CCA does not own the orthogonality property, which is crucial for information reconstruction. In this paper, we utilize an orthogonal variant of CCA, which has been proven by experiments to achieve a better performance than the original CCA in terms of global face reconstruction.