Convolutional neural networks based image resampling with noisy training set
Andrey V. Nasonov, Konstantin Chesnakov, Andrey Serdjevich Krylov · 2016
A new learning model for image resampling with convolutional neural network is proposed. Its main idea is the dataset preparation method for deep learning. The proposed algorithm can work with noisy and noiseless images and provides good quality for wide noise level range. The method was tested using standard datasets and was also applied for retinal image resampling.