Swarm Filtering Procedure and Application to MRI Mammography

摘要

Abstract—Research on swarming has primarily focused on applying swarming behavior with physics-derived or ad-hoc models to tasks requiring collective intelligence in robotics and optimization. In contrast, applications in signal processing are still lacking. The purpose of this paper is to investigate the use of biologically-inspired swarm methods for signal filtering. The signal, in the case of images the grayscale value of the pixels along a line in the image, is modeled by the trajectory of an agent playing the role of the prey for a swarm of hunting agents. The swarm hunting the prey is the system performing the signal processing. The movement of the center of mass of the swarm represents the filtered signal. The position of the center of mass of the swarm during the virtual hunt is reverted into grayscale values and represents the output signal. We show results of applying the swarm-based signal processing method to MRI mammographies. Index Terms—Swarm intelligence, nonlinear signal filter, MRI, mammography, image processing. I.