Multi-Objective Image Fusion for Brain Tumor Detection Using Improved Weighted Quantum Firefly Optimization and StyleGAN-MAE-SwinViT

To improve diagnostic precision, the accurate fusion of imaging methods is necessary for brain tumor identification from imaging studies.Conventional fusion techniques frequently encounter issues such as noise interference, low contrast, and data loss, which reduce their effectiveness in clinical settings.This paper proposes a Multi-Objective Image Fusion architecture that combines StyleGAN-MAE-ViT and Improved Weighted Quantum Firefly Optimization (IWQFO) to address these challenges.The IWQFO method employs a quantum-inspired searching process to balance multiple objectives, including brightness enhancement, edge preservation, and architectural resemblance, to optimize the fusion process.Meanwhile, StyleGAN-MAE-ViT integrates the advantages of the Vision Transformer (ViT) for spatial attention-based tumor segmentation, the Masked Autoencoder (MAE) for robust feature reconstruction, and StyleGAN for high-fidelity image generation.To preserve critical tumor information while eliminating redundant noise, the proposed architecture fuses multi-modal MRI images (T1, T2, and FLAIR).Experimental evaluations conducted on benchmark brain tumor datasets demonstrate that the proposed approach outperforms existing fusion techniques in terms of Peak Signal-to-Noise Ratio (PSNR), tumor segmentation accuracy, Feature Similarity Index (FSIM), and Structural Similarity Index (SSIM).These findings validate the superiority of the IWQFO-StyleGAN-MAE-ViT fusion model in enhancing tumor visibility, aiding radiologists in making accurate and timely diagnoses.