Autoencoder Network based CT and MRI Medical Image Fusion

In the last twenty years, medical imaging has played a vital role in the diagnosis of diseases, but relying on a single imaging modality often provides insufficient information. Image fusion, which combines images from different medical devices, is essential for comprehensive analysis. This research focuses on developing an advanced system that integrates autoencoder and nuclear norm techniques to achieve optimal fusion. The method utilizes nuclear norm and residual networking to merge complementary information obtained from diverse imaging techniques. Residual interconnections connect the results of convolutional layers, and the nuclear norm technique combines image features, resulting in a fused image. Evaluation metrics, including structural similarity index measure, mutual information, and Qabf values (0.6672,4.6223, and 0.7989), checks the efficiency of the fusion algorithm. Compared to existing techniques, this approach demonstrates superior fusion outcomes, effectively combining CT and MRI images and helpful in the diagnosis of brain cancer.