IoT-Enhanced GAN Framework for Multi-Modal Medical Image Enhancement with Residual Attention and Fusion

Abstract

Medical image enhancement is essential for improving diagnostic image quality and supporting clinical decision-making. With the increasing integration of the Industrial Internet of Things (IoT) in healthcare, real-time data from medical devices and sensors has become a key factor in enhancing multimodal medical images. In this paper, we propose a novel GAN-based model that integrates a residual attention mechanism, multi-scale convolutional networks, and adaptive upsampling modules. The model effectively enhances both local details and global structures in multimodal medical images, addressing the limitations of existing methods. We introduce a dual-discriminator design consisting of global and PatchGAN discriminators, to ensure the authenticity and quality of the generated images. Additionally, the integration of IoT data, such as real-time physiological monitoring, enables the model to dynamically adjust parameters for personalized and context-aware image enhancement. Experimental results on the BraTS and LIDC-IDRI datasets demonstrate that our model outperforms existing methods in terms of PSNR, SSIM, and visual quality. The results indicate that the model preserves fine details and maintains structural consistency, highlighting its potential for medical image enhancement. These findings suggest that the proposed approach could serve as a reliable solution for clinical imaging tasks, where high-quality image enhancement is critical for accurate diagnosis and treatment planning.