Self-Adaptive PCNN Based on the ACO Algorithm and its Application on Medical Image Segmentation

Medical image segmentation plays a dominant role in medical image analysis and clinical research. As an effective method for image segmentation, pulse-coupled neural networks (PCNN) has its own limitation that the values of the parameters have a strong effect on its performance when it is used to segment the image. This paper proposed a new method for medical image segmentation using the self-adaptive PCNN model. In this method, we combined the searching capabilities of ant colony optimization (ACO) algorithm in the solution space with the biological characteristics of PCNN, to find the optimal values of PCNN's parameters for each input image. Moreover, the search process of the ACO algorithm was divided into the local searching and the global searching to accelerate the speed of the ASO's convergence. Based on the above work, a new automatic method for the image segmentation, namely ACO-PCNN, was presented. Lastly, four pairs of different MR medical images, including transaxial, sagittal, coronal sections and noisy medical images, were used to test and validate the performance of the proposed method. The experimental results illustrated that our method was accurate and effective to MRI medical images.