Super-resolved segmentation in blurred and noisy images by iterative edge reconstruction

We present ITEROOF, a novel, to the best our knowledge, method that enables the segmentation of objects separated by distances below the imaging system's resolution limit and therefore appear overlapped in the images. The method is based on an iterative approach to a recently proposed algorithm for edge detection with super-resolution. We characterized its accuracy and precision in simulated and real images, performing different signal-to-noise ratios, showing that the method can achieve segmentation with an edge location accuracy better than one-fifth of the image resolution, even in very demanding noise conditions.