Research Area:  Machine Learning
In recent times the use of digital images has increased the demand for high-resolution images. The images captured are sometimes affected by noise, making visualization of the objects difficult, so the image super-resolution method is used to solve this problem. This research is based on a predefined Edge Informed Single Image Super-Resolution(EISR). The model is based on a deep learning approach that uses a convolutional neural network(CNN) and works on single image super-resolution(SISR). The first stage of the proposed model is the bi-cubic interpolation stage, followed by the Edge enhancement and Image completion stage. A qualitative comparison between the existing and proposed models on the x2 scaling factor is made.
Keywords:  
Author(s) Name:   Aniket Zope; Vandana Inamdar
Journal name:  
Conferrence name:  2nd Global Conference for Advancement in Technology (GCAT)
Publisher name:  IEEE
DOI:  10.1109/GCAT52182.2021.9587565
Volume Information:  
Paper Link:   https://ieeexplore.ieee.org/abstract/document/9587565