This thesis aims to advance research in image segmentation by developing robust techniques for evaluating image segmentation algorithms. The key contributions of this work are as follows. First, we investigate the characteristics of existing measures for supervised evaluation of automatic image segmentation algorithms. We show which of these measures is most effective at distinguishing perceptually accurate image segmentation from inaccurate segmentation. We then apply these measures to evaluating four state-of-the-art automatic image segmentation algorithms, and establish which best emulates human perceptual grouping. Second, we develop a complete framework for evaluating interactive segmentation algorithms by means of user experiments. Our system comprises evaluation measures, ground truth data, and implementation software. We validate our proposed measures by showing their correlation with perceived accuracy. We then use our framework to evaluate four popular interactive segmentation algorithms, and demonstrate their performance. Finally, acknowledging that user experiments are sometimes prohibitive in practice, we propose a method of evaluating interactive segmentation by algorithmically simulating the user interactions. We explore four strategies for this simulation, and demonstrate that the best of these produces results very similar to those from the user experiments.