WITHDRAWN: Deep learning based image processing approaches for image deblurring

Abstract

Mobile phone imagery has evolved substantially over the past two decades. The camera is one of the primary features of a new cell phone. A lot of research in this field has been done to enhance the quality of the image. A small phone structure limits the camera module mounted in cell phones, which means that there can be no mounted thick lenses or large image sensors on the phones. This affects the volume and the picture quality of the captured light. With less light, phones perform worse than digital Single-Lens Reflex (DSLR) cameras as opposed to image quality tests. Post processing is then used to enhance the quality of the image. Dark conditions and fast movement in mobile imaging are difficult. In dark conditions, longer exposure period collects more light, which can cause movement blurred objects. Motion blur artefact, e.g. when photo to graphing a racing car, and may also be caused by fast moving object at daylight. The movement blur causes sharp information to be lost and thus poor picture quality. A deblur ring is the tool used to eliminate flutter from photographs and make them appear clearer. In the field of signal processing deep learning-based approaches have recently become popular. The results have been promising since profound learning algorithms can learning and model nonlinear and complex connections. Deep learning algorithms were also used for several tasks in image restore work, as was done here.