Abstract:
This research addresses the challenge of blurry and low-resolution images, which often lack the necessary
information for effective user perception. Blurriness in images can occur due to various factors such as camera movement,
improper focus, aperture settings, and external influences, resulting in degraded or deteriorated photographs. Additionally, the
presence of haze, whether uniform or non-uniform, can further contribute to image distortion and poor quality. To tackle these
issues, a deep learning approach utilizing convolutional neural networks (CNNs) is proposed. This approach simultaneously
addresses image de-blurring and super resolution, providing a comprehensive solution. By employing super resolution
techniques, it becomes possible to enhance the quality of images significantly, generating high-resolution outputs from low resolution inputs. The use of neural networks, particularly CNNs, demonstrates experimental superiority over existing deep
learning algorithms, leveraging the benefits of super resolution. The suggested model exhibits scientific evidence of the
effectiveness and efficiency of the proposed system. It demonstrates that the quality and quantity of the system's performance
are achieved. Regardless of the level of blur in the input images, the proposed model can achieve high-quality resolution,
surpassing the limitations of current approaches. By employing a profound learning strategy through CNNs, both known and
unknown levels of blur can be effectively addressed, resulting in superior image restoration and enhancement.