Learning-based Hierarchical Clustering Regression for Efficient Unaligned Face Hallucination

Abstract

In this paper, we propose an effective single image super-resolution method for unaligned face images, in which the learning-based hierarchical clustering regression approach is used to get better reconstruction model. The proposed face hallucination method can be divided into two parts: clustering and regression. In the clustering part, a dictionary is trained on the whole face image with tiny size, and the training images are clustered based on the Euclidean distance. Thus, the facial structural prior is fully utilized and the accurate result of clustering can be obtained. In the regression part, only one global dictionary in which atoms are taken as the anchors, will be trained in the entire training phase. Therefore, the time complexity can be effectively reduced. More importantly, the learned anchors are shared with all the clusters. For each cluster, the Euclidean distance is used to search the nearest neighbors for each anchor to form the subspace. Moreover, a regression model is learned to map the relationship between low-resolution features and high-resolution samples in every subspace. The core idea of our method is to utilize the same anchors but different samples for clusters to learn the local mapping more accurately, which can reduce training time and improve reconstruction quality. Experimental results show that the proposed method outperforms some state-of-the-art methods.