Fourier-CPPNs for Image Synthesis
This work addresses a limitation in neural art generation for creative applications, but it is incremental as it builds on existing CPPN methods.
The authors tackled the problem of CPPNs generating overly smooth images lacking high-frequency detail by extending them to explicitly model frequency information, resulting in improved visual detail for image synthesis.
Compositional Pattern Producing Networks (CPPNs) are differentiable networks that independently map (x, y) pixel coordinates to (r, g, b) colour values. Recently, CPPNs have been used for creating interesting imagery for creative purposes, e.g., neural art. However their architecture biases generated images to be overly smooth, lacking high-frequency detail. In this work, we extend CPPNs to explicitly model the frequency information for each pixel output, capturing frequencies beyond the DC component. We show that our Fourier-CPPNs (F-CPPNs) provide improved visual detail for image synthesis.