NORMA eResearch @NCI Library

Single Image Super Resolution Using Multiple Deep Convolution Neural Network

Rajpara, Khushbu Mukeshbhai (2021) Single Image Super Resolution Using Multiple Deep Convolution Neural Network. Masters thesis, Dublin, National College of Ireland.

[thumbnail of Master of Science]
PDF (Master of Science)
Download (839kB) | Preview
[thumbnail of Configuration manual]
PDF (Configuration manual)
Download (1MB) | Preview


Image super-resolution has long been the subject of controversy in the areas of computer vision and image processing. Deep learning has contributed significant advances in the past several decades in achieving high-resolution images with additional information utilizing deep neural networks. Modern super-resolution (SR) techniques often use convolutional neural networks to read complex non-redirected maps between paired LR and HR images. This study will be focusing on deep neural network usage methods for image enhancing activity. High-resolution images are utilized as the target while degraded images are used as the network's input to get desired outcomes. Convolution neural network (CNN), Auto Encoder model, Multiscale learning model, and enhanced deep residual network (EDSR) are the four deep neural network models employed in this study. To assess the evaluation criteria and gain a comparative analysis among the models. The models are experimenting on public datasets and attempting to use a GPU infrastructure with minimal training times. The efficiency of the model is demonstrated by a comparative analysis of the models which is based on a test image dataset using RMSE, the accuracy of the model, and PSNR metrics. The accuracy of the Auto Encoder model is the highest among the four networks, according to the findings of testing. With 30 epochs, the Autoencoder has an accuracy of 86.37 % and an MSE of 0.0177 with the highest PSNR value which is 34.56 dB.

Item Type: Thesis (Masters)
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
T Technology > TR Photography
Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: Tamara Malone
Date Deposited: 01 Mar 2023 12:39
Last Modified: 01 Mar 2023 17:29

Actions (login required)

View Item View Item