Denoising in the Dark: Privacy Preserving Deep Neural Network Based Image Denoising

Abstract

          Large volumes of images are being exponentially generated today from various kinds of imaging devices (smart phones, medical imaging equipment, digital cameras, etc.). Such tremendous growth is further greatly accelerated along with the fast development and wide deployment of diverse Internet of Things (IoT) applications. In the existing system, a secure cloud-based image service framework is presented, which allows privacy-preserving and effective image denoising on the cloud side to produce high-quality image content, a key for assuring the quality of various image-centric applications.

          We resort to state-of-the-art image denoising techniques based on deep neural networks (DNNs), and show how to uniquely bridge cryptographic techniques (like lightweight secret sharing and garbled circuits) and image denoising in depth to support privacy-preserving DNN based image denoising services on the cloud. By design, the image content and the DNN model are all kept private along the whole cloud-based service flow. In the proposed system, to ensure reliable and secure communication, design of efficient image denoising scheme, with classification model using Deep learning algorithms enrolled together to form a new algorithm named as Deep-Dark-Net is evaluated. The benefit of proposed system is to ensure added security in Dark Cloud using newly structured deep learning algorithm.

Proposed System

          In the proposed system, to ensure reliable and secure communication, design of efficient image denoising scheme, with classification model using Deep learning algorithms enrolled together to form a new algorithm named as Deep-Dark-Net is evaluated. The benefit of proposed system is to ensure added security in Dark Cloud using newly structured deep learning algorithm.