BDCN和U-net边缘生成两阶段修复算法

为了对图像中大面积不规则缺失区域进行合理的结构修复和精细的纹理重构，提出了一种基于双向级联边缘检测网络（BDCN）和U-net残缺边缘生成的两阶段网络图像修复算法.第一阶段首先基于BDCN网络提取边缘，然后基于U-net架构的边缘生成网络用下采样对缺失图像边缘提取特征，再结合上采样的输入信息和下采样各层信息还原图像边缘纹理细节；第二阶段使用空洞卷积进行下采样和上采样，经过残差网络重建细节丰富的缺失图像.在公开数据集上将所提算法与现有经典算法进行对比，实验结果表明，所提算法能得到合理的结构和精细的纹理细节，优于对比算法.

Two-Stage Network Inpainting Algorithm Based on BDCN and U-net Edge Generation

To repair the large irregular missing areas of an image, and obtain reasonable structure and fine-detailed textures,a two-stage network image inpainting algorithm based on bi-directional cascade network (BDCN) for perceptual edge detection and U-net incomplete edge generation is proposed. Firstly,the algorithm edges are extracted using the BDCN network. In the first stage,down-sampling is used to extract the features of the missing image edges based on the edge generation network of the U-net network architecture. The information inputted by the up-sampling layer and the information of each down-sampling layer is then combined to restore the image edge texture details. In the second stage,the hole convolution is applied for down-sampling and up-sampling,which adopts the residual network to reconstruct the missing image with rich details. The proposed algorithm is compared with the existing classic algorithm on the public datasets. Experimental results demonstrate that the proposed algorithm can obtain reasonable results and fine texture details,and its performance is superior to those of the contrast algorithms.