GIFnets: Differentiable GIF Encoding Framework — mentioned by dbaupp

Wed, 16 Sep 2020 10:47:26 GMT

Very interesting! This seems somewhat similar to the recently published GIFnets[1]. However, I believe GIFnets is training a reusable network to a predict palettes, and pixel assignments, while this post is focusing on optimising the "weights" (i.e. pixel values) for a single image. I wonder if the loss functions from GIFnets could be applied to this single-image approach to potentially solve the banding problem via something a little more "perceptual" than the variance term mentioned. [1]: "GIFnets: Differentiable GIF Encoding Framework" https://arxiv.org/abs/2006.13434

Paper and discussion history. A mention is not an endorsement.

GIFnets: Differentiable GIF Encoding Framework · HackerPapers discussions

GIFnets: Differentiable GIF Encoding Framework — mentioned by dbaupp