Using checkerboard rendering and deconvolution to eliminate checkerboard artifacts in images generated by neural networks

The images generated by deep neural networks is clear, but when the observers watch very closely at these images, they often see some checkerboard patterns of artifacts, in this paper, we analyzed the causes of this phenomenon of checkerboard artifacts, and we have found a solution to the problem, deconvolution and checkerboard rendering can provide a method to eliminate the checkerboard artifacts and make the images more distinct. Experimental results show that we have provided to use solution that improves the quality of many approaches to generating images with deep neural networks.

[1]  Luc Van Gool,et al.  The Pascal Visual Object Classes Challenge: A Retrospective , 2014, International Journal of Computer Vision.

[2]  Christian Ledig,et al.  Is the deconvolution layer the same as a convolutional layer? , 2016, ArXiv.

[3]  Daniel Rueckert,et al.  Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).