<p>The quality of images and accuracy in diagnosis can be significantly impacted by noise in the medical images. Noise can interfere with diagnostic tasks, such as describing object boundaries, which must be performed for accurate diagnosis. Medical imaging specialists may significantly enhance image quality, smaller errors, and improve the precision of evaluation and therapy by utilizing denoising techniques. In the modern healthcare system, imaging technology is crucial for precise diagnosis and effective treatment plans. However, a variety of sources of noise frequently impair the appearance and comprehension of medical images. This research develops a novel work model for denoising the medical image. It has the ability to remove the noise in the medical images, which significantly increases diagnostic accuracy. Here, the innovative model named as Region-vision transformer-based adaptive mobile-UNet +  + with Kalman filter (RViT-AMUNet-KF) is developed for noise removal in the images. The RViT-AMUNet-KF has less complexity and provides better results in the medical image denoising process. From the Kalman filter, the gradient information and noise are utilized by the RViT-AMUNet model to get better results in removing the noise. Additionally, the incorporation of the Kalman filters with the RViT-AMUNet images. Further, the attributes of the MobileUNet +  + are tuned by utilizing the Decision Variable Updated Far and Near Optimization. The experimental evaluation is executed to maximize the effectiveness of the designed denoising model. The proposed model attained PSNR of 58.52, 60.0, MSE of 0.09, 0.06, RMSE of 0.30, 0.25, and SSIM of 0.99, 0.99 on datasets 1 and 2.</p>