Now a day’s medical images, especially images of internal organs taken by the aid of X-rays, CT scans and
MRI need large space for storing and the compression technique need to be introduced to diminish the
storage space consumed by cloud. The vital information of the medical images needs to be compressed
without distorting details. Also the ubiquitous noise has to be reduced with less computation from the
medical images for further processing. Manual reading of a medical report lacks the accuracy of finding
exact illness. Various Deep Learning techniques like Convolutional Neural Network, Recurrent Neural
Network, Autoencoders and Generative Adversarial Network architectures made the diagnosis easily and
accurately. The experiment conducted on denoising on brain tumor dataset using auto encoder is presented
in this paper and which can be further used for applications of medical imaging. A tensor flow based frame
work is proposed to denoise Magnetic Resonance Image with minimum loss of information.