Time-weighted evaluation of image segmentation with a genetic algorithm

The performance of a segmentation algorithm can be evaluated by systematic comparison with hand-segmented ground-truth images. When evaluation extends over an algorithm's parameter space, then the search for satisfactory settings has a considerable cost in time. This paper considers applying a genetic algorithm (GA) to avoid an exhaustive search. To further reduce evaluation time and subsequent image batch-processing times, this paper introduces a time factor into the GA cost function. This procedure while preserving the GA solution, selection of parameters to minimize the fit to hand-segmented images, also improves interpretation and parameter selection.