The field of image processing has paid a lot more attention to pixel-level picture fusion over the past two decades. Taking an average of each pixel in the source photos is the simplest method for fusing images at the pixel level. In, we see an example of this type of method where two medical images are averaged together to get a single coherent image. While the averaging approach is straightforward to put into practise, it has a number of downsides, including decreased contrast that can result in a significant defeat of data. The authors of this manuscript report deriving a novel fusion model in terms of an unsupervised classification model called principle component analysis, which takes into account the structure and sparse restrictions present in multi-modality images. Eigen value and Eigen vector correlation and covariance analyses make image fusion more adaptable and precise. CT images are ideal for studying rigid objects like bones and implants because of the minimal distortion they introduce, but CT scans are incapable of picking up physiological shifts. Soft tissue can be more clearly seen in an MR image. To measure the effectiveness of the proposed workbench, its performance is simulated in MATLAB for state-of-the-art methods. Key metrics include PSNR (peak signal-to-noise ratio), MSE (Mean square error), association, and precision. All metrics point to the suggested workbench's superior performance, and it also reduces computational load.