Abstract—The implementation of operators known as edge detection in medical images at times lead to elicitation of unimportant or inaccurate information as to the target spot in the image. In the compressed images and the ones with noise, edge detection accuracy reduces. In this paper, the technique of edge detection of medical images based on ant colony algorithm is introduced. By dispatching ants to the image pixels and relying on edge specifications, the pheromones matrix is formed which contains information related to the damaged tissue. Receiving 220 medical images composed of 90 retina images taken from diabetic patients, 80 MRI images as well as 50 microscopic images taken from various medical databases and applying system to them in contrast to such known operators as Canny and Sobel, an acceptable level of accuracy 94.90%, sensitivity 94.16% and specificity 94% was separated in the target area from the rest of image. The 88.79% Kappa coefficient indicates the high reliability factor of system in terms of performance. The application of this system to tissue imaging systems not only increases the accuracy of detection, but also steps up the process speed to a large extent. Reduced expenses, cost savings in long term and non-destructive quality of this system are the main distinguishing features of this system.

Index Terms—edge detection, image processing, ant colony, medical images, cancerous masses


Cite: Mohammad Rezaee and Mohammad Bagher Tavakoli, "Cancerous Masses Segmentation by Using Heuristic Ant Colony Algorithms in Medical Images," Journal of Image and Graphics, Vol. 2, No. 2, pp. 128-134, December 2014. doi: 10.12720/joig.2.2.128-134