Abstract

Image segmentation is a key component in many com-puter vision systems, and it is recovering a prominent spotin the literature as methods improve and overcome theirlimitations. The outputs of most recent algorithms are in theform of a hierarchical segmentation, which provides seg-mentation at different scales in a single tree-like structure. Commonly, these hierarchical methods start from some lowlevel features, and are not aware of the scale information of the different regions in them. As such, one might need towork on many different levels of the hierarchy to find theobjects in the scene. This work tries to modify the existing hierarchical algorithm by improving their alignment, that is, by trying to modify the depth of the regions in the tree to better couple depth and scale. To do so, we first train a regressor to predict the scale of regions using mid-level features. We then define the anchor slice as the set of regions that better balance between over-segmentation and undersegmentation. The output of our method is an improved hierarchy, re-aligned by the anchor slice. To demonstrate the power of our method, we perform comprehensive experiments, which show that our method, as a post-processing step, can significantly improve the quality of the hierarchical segmentation representations, and ease the usage of hierarchical image segmentation to high-level vision tasks such as object segmentation. We also prove that the improvement generalizes well across different algorithms and datasets, with a low computational cost.1