Creating photorealistic images with neural networks and a Gameboy Camera
pinchofintelligence.com
Creating photorealistic images with neural networks and a Gameboy Camera
1–10 of 36 posts
Re: Creating photorealistic images with neural networks and a Gameboy Camera
#2So, smoothing and applying a brownish/pink skin color.
Re: Creating photorealistic images with neural networks and a Gameboy Camera
#3If this keeps up, that trope of CSI's "enhance, zoom, enhance" will become reality.
Re: Creating photorealistic images with neural networks and a Gameboy Camera
#4So, smoothing and applying a brownish/pink skin color.
The model clearly overfits if to compare the fidelity of pictures from the training set vs the out of sample ones (sent in by others).
Re: Creating photorealistic images with neural networks and a Gameboy Camera
#5"Thanks to this post http://distill.pub/2016/deconv-checkerboard/ I found out about the alternative deconvolution layer and implemented that one."
Yeah, I saw the same thing with WGAN. ConvTranspose2d is not that great for upscaling because it creates artifacts. That said, the post actually recommends doing a 'subpixel' convolution for upscaling (something like, do a convolution out to channels for each pixel, then use PixelShuffle to map it back into a 3-channel image), not doing a bilinear/nearest-neighbor + Conv2d(3,3).
(A GAN would probably also deliver better colorizing results in general.)
Re: Creating photorealistic images with neural networks and a Gameboy Camera
#6So, smoothing and applying a brownish/pink skin color.
That's pretty much what I thought. I think the concept was pretty neat, and it does a pretty job of smoothing the gradients while maintaining the detail, and I think that's probably the hardest part.
It would be hard for it to accurately get skin color, since the camera is only seeing skin, and it all gets leveled to a similar lightness. So, I'm not surprised that skin color was hard, but the report still probably should not have included the line "Note that even skincolor is accurate most of the times" -- I guess "most" could almost be considered accurate, if the majority of the samples had the same pinkish skin to start with. Almost all the results from the celebrity dataset ended up with the exact same tone, despite wildly different input tones: http://imgur.com/a/daJUa
Re: Creating photorealistic images with neural networks and a Gameboy Camera
#7How fast does it take to run? could it be done in real time for compression? Also, wonder how low of resolution you could go?
Re: Creating photorealistic images with neural networks and a Gameboy Camera
#8So, smoothing and applying a brownish/pink skin color.
What would be really interesting to see in these kinds of articles is how well you could do with just a batch Photoshop operation hand-tuned on a handful of photos, and then a comparison of how much better the NN does.
Re: Creating photorealistic images with neural networks and a Gameboy Camera
#9If this keeps up, that trope of CSI's "enhance, zoom, enhance" will become reality.
Except instead of enhance you have "Make up something based on your training set".