Towards efficient medical lesion image super-resolution based on deep residual networks

摘要

Super-resolution reconstruction helps doctors clearly observe the details of medical lesion images and increases the likelihood of the disease being diagnosed and cured. In this paper, we propose an efficient medical lesion image super-resolution method based on deep residual networks. First, a multi-scale super-resolution reconstruction model based on a deep residual network is trained. Second, an easy-to-use interface is designed. Third, a multi-scale super-resolution reconstruction model is used to reconstruct different types of medical lesion images with different scales and calculate their peak signal-to-noise ratios and structural similarity index values; The experimental results show that the proposed super-resolution reconstruction method achieves superior performance over the other methods compared in this work.