CT kernel conversions using convolutional neural net for super-resolution with simplified squeeze-and-excitation blocks and progressive learning among smooth and sharp kernels.
Da-in EunIlsang WooBeomHee ParkNamkug KimSang Min LeeJoon Beom SeoPublished in: Comput. Methods Programs Biomed. (2020)
Keyphrases
- neural nets
- super resolution
- image reconstruction
- high resolution
- low resolution
- kernel function
- single image super resolution
- kernel methods
- learning tasks
- learning algorithm
- back propagation
- feed forward
- high quality
- image super resolution
- supervised learning
- learning process
- image restoration
- prior knowledge
- active learning
- multi frame
- neural network
- unsupervised learning
- artificial neural networks
- sparse coding
- dictionary learning
- high resolution images
- resolution enhancement