Nearest feature line embedding for face hallucination

A new manifold learning method, called nearest feature line (NFL) embedding, for face hallucination is proposed. While many manifold learning based face hallucination algorithms have been proposed in recent years, most of them apply the conventional nearest neighbour metric to derive the subspace and may not effectively characterise the geometrical information of the samples, especially when the number of training samples is limited. This reported work proposes using the NFL metric to define the neighbourhood relations between face samples to improve the expressing power of the given training samples for reconstruction. The algorithm preserves the linear relationship in a smaller local space than traditional manifold learning based methods, which better reflects the nature of manifold learning theory. Experimental results demonstrate that the method is effective at preserving detailed visual information.