Augmented reality on mobile phones has recently made major progress. Lightweight, markerless object recognition and tracking makes handheld Augmented Reality feasible for new application domains. As this field is technology driven the interface design has mostly been neglected. In this paper we investigate visualization techniques for augmenting printed documents using handheld Augmented Reality. We selected the augmentation of printed photobooks as our application domain because photobooks are enduring artefacts that often have online galleries containing further information as digital counterpart. Based on an initial study, we designed two augmentations and three techniques to select regions in photos. In an experiment, we compare an augmentation that is aligned to the phone's display with an augmentation aligned to the physical object. We conclude that an object aligned presentation is more usable. For selecting regions we show that participants are more satisfied using simple touch input compared to Augmented Reality based input techniques.