Deconvolution and Checkerboard Artifacts
AUGUSTUS ODENA Google Brain
VINCENT DUMOULIN Université de Montréal
CHRIS OLAH Google Brain
Oct. 17 2016
Citation: Odena, et al., 2016
When we look very closely at images generated by neural networks, we often see a strange checkerboard
pattern of artifacts. It’s more obvious in some cases than others, but a large fraction of recent models exhibit
this behavior.
Mysteriously, the checkerboard pattern tends to be most prominent in images with strong colors. What’s
going on? Do neural networks hate bright colors? The actual cause of these artifacts is actually remarkably
simple, as is a method for avoiding them.
Deconvolution & Overlap
When we have neural networks generate images, we often have them build them up from low resolution,
high-level descriptions. This allows the network to describe the rough image and then fill in the details.
In order to do this, we need some way to go from a lower resolution image to a higher one. We generally do
this with the deconvolution operation. Roughly, deconvolution layers allow the model to use every point in
the small image to “paint” a square in the larger one.
Radford, et al., 2015 Salimans et al., 2016 Donahue, et al., 2016 Dumoulin, et al., 2016
[1] [2] [3] [4]