A metric for evaluating and comparing hierarchical and multi-scale image segmentations

We present a method for evaluating the quality of a multi-scale or hierarchical image segmentation against a reference single-scale segmentation, for comparing different seg- mentation algorithms or fine-tuning an algorithm's parameters to a specific application. I. INTRODUCTION Image segmentation is usually performed as a preprocessing step for many image understanding applications, for example in some land-cover and land-use classification systems. A segmentation algorithm is used with the expectation that it