Text this: Evaluating supervised learning approaches for spatial-domain multi-focus image fusion