A Lung Cancer Tumor Image Segmentation Method of A SFC-MSPCNN based on PET/CT

To enhance the low segmentation accuracy and computational complexity of traditional pulse coupled neural network (PCNN) in medical image processing, a Synchronous FC-MSPCNN (SFC-MSPCNN) is proposed to realize lung cancer mass image segmentation. Compared with previous PCNN models, this method further optimizes and improves the weight matrix Wijkl, link strength, and dynamic threshold amplitude, simplifies the setup parameters, and reduces the number of iterations within the effective transmission period. Additionally, we have added a balance parameter K to adjust the dynamic threshold. A large number of related experiments demonstrate that our method has better results compared to other algorithms, which can accurately segment lung cancer masses, and significantly reduces the randomness and unpredictability of firing neurons.