The High-Resolution chest CT scan images help to diagnose lung related diseases accurately. In general, the more advanced hardware used in CT Scan machines, the more high resolution images will be generated. But it is a costlier approach. This limitation can be overcome with the post processing of the images generated from the CT machine. Even when the image is upscaled, the quality of the image should be retained. So, the process of reconstructing the High-Resolution images from the Low-Resolution images is known as Image Super-Resolution. The recent advancements in hardware and Super Resolution deep neural networks enabled reconstructing High-Resolution images in an efficient way. The objective quality metric Peak-Signal-to-Noise-Ratio evaluates the performance of a SR deep model. In this paper, proposed a pixel attention based deep neural network, MediSR for chest CT scan medical image Super-Resolution. The model is trained with two chest CT datasets and the experimental results showed an improvement of 1.78% and 18.23% for the 2 × and 4 × scale factors over the existing literature.
CITATION STYLE
Rajeshwari, P., & Shyamala, K. (2023). Pixel Attention Based Deep Neural Network for Chest CT Image Super Resolution. In Communications in Computer and Information Science (Vol. 1798 CCIS, pp. 393–407). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-28183-9_28
Mendeley helps you to discover research relevant for your work.