Single Image Super-Resolution (SISR) aims to generate a high-resolution (HR) image of a given low-resolution (LR) image. The most of existing convolutional neural network (CNN) based SISR methods usually take an assum...
详细信息
ISBN:
(纸本)9781450371896
Single Image Super-Resolution (SISR) aims to generate a high-resolution (HR) image of a given low-resolution (LR) image. The most of existing convolutional neural network (CNN) based SISR methods usually take an assumption that a LR image is only bicubicly down-sampled version of an HR image. However, the true degradation (i.e. the LR image is a bicubicly downsampled, blurred and noisy version of an HR image) of a LR image goes beyond the widely used bicubic assumption, which makes the SISR problem highly ill-posed nature of inverse problems. To address this issue, we propose a deep SISR network that works for blur kernels of different sizes, and different noise levels in an unified residual CNN-based denoiser network, which significantly improves a practical CNN-based super-resolver for real applications. Extensive experimental results on synthetic LR datasets and real images demonstrate that our proposed method not only can produce better results on more realistic degradation but also computational efficient to practical SISR applications.
During the last decade, finite element (FE) modelling has become ubiquitous in understanding complex mechanobiological phenomena, e.g. bone-implant interactions. The extensive computational effort required to achieve ...
详细信息
During the last decade, finite element (FE) modelling has become ubiquitous in understanding complex mechanobiological phenomena, e.g. bone-implant interactions. The extensive computational effort required to achieve biorealistic results when modelling the post-yield behaviour of microstructures like cancellous bone is a major limitation of these techniques. This study describes the anisotropic biomechanical response of cancellous bone through stress-strain curves of equivalent bulk geometries. A cancellous bone segment, reverse engineered by micro computed tomography, was subjected to uniaxial compression. The material's constitutive law, obtained by nano-indentations, was considered during the simulation of the experimental process. A homodimensionally bulk geometry was employed to determine equivalent properties, resulting in a similar anisotropic response to the trabecular structure. The experimental verification of our model sustained that the obtained stress-strain curves can adequately reflect the post-yield behaviour of the sample. The introduced approach facilitates the consideration of nonlinearity and anisotropy of the tissue, while reducing the geometrical complexity of the model to a minimum.
暂无评论