Image reconstruction system and method combining local and global information

The invention belongs to the field of image processing, and relates to an image reconstruction system and method combining local and global information, and the system comprises a generator network, an edge detection network and a depth detection network. The generator network maps the input image Y with partial information loss to obtain a reconstructed image Z; an end-to-end edge detection network is adopted, and an edge image is directly generated as output after an image is input; a depth detection network for extracting a single image is adopted, and the pixel-level depth is directly predicted after the whole image is input; compared with a mask self-encoding model based on image reconstruction in the prior art, the network architecture provided by the invention realizes the optimal balance between the image reconstruction accuracy and the calculation efficiency, and the accuracy and the calculation efficiency of image reconstruction can be improved. 本发明属于图像处理领域，涉及一种联合局部和全局信息的图像重建系统及方法，所述系统包括生成器网络、边缘检测网络