Zebras, Horses & CycleGAN – Computerphile
Articles Blog

Zebras, Horses & CycleGAN – Computerphile

August 12, 2019

Only registered users can comment.

  1. I don't care if I'm first or second or anything like that, just wanted to use this opportunity to tell that this channel is amazing!

  2. Hello friend, you can answer my question and the doubt of many subscribers about java script and php. Please 🙏

  3. Great video ! Especially the discussion about how reliable or trustworthy super-resolution can be when using a single image.

    If the information is not captured by the camera, there's nothing you can do about it. You need some extra knowledge about what you're looking at, like for instance more low resolution pictures.

  4. I was expecting you to reveal at the end of the video that the styling of your shirt was being CycleGAN-generated all along.

  5. I get you guys are required by law to say Zed instead of Zee, but did this dude REALLY just say Zedbra for Zebra? omg.

  6. basically you are assuring that F (and in return also G) are bijections, i.e. they don't "collapse" the domains

  7. My favourite CycleGAN fail story is one that was made to take areal photos and turn into maps. It made nonsense maps and was suspiciously good at recreating lamp posts, fans on top of buildings, and so on.

    Turns out it did make nonsense maps, then hid information about the original image in the least significant bits of the pixels in those maps.

    I don't remember where I read about it, though. So you might have to take its truth with a grain of salt.

  8. If one were to convert to a monet painting then it would be lossy so wouldn't a cyclegan incentivise it to be a bit less like a monet so that there is enough data to covert it back to the original?

    If the network gives you the same zebra every time but places the original, shrunk down and in the corner then the reverse network can just take that corner and expand it? Does that ever happen?

  9. I saw NASA pull out a licence plate number from a few pixels by tracking the differences between frames on a movie and back calculating what the image must have been to produce that.
    Maybe combining that idea with this one can get a better 2001.

  10. It would be interesting to see to what degree a cycleGAN can be trained to make interesting videos like this one from a scientific paper by having to reproduce a matching paper from the video.

  11. I need to get into this kind of thing.
    I'd love to build one to remove telephone/power lines from my photos.

  12. Honestly I am not convinced that sort of enhancement is going to ever really be applicable for legal purposes precisely because it is always going to be guesswork at least unless you are dealing with an analogue source that does in fact have fine detail that is just too difficult for humans to make out. That said I can potentially see one are these could eventually find use in police work and that is perhaps as a far cheaper and thus more accessible way to go about creating images based on eyewitness descriptions etc, currently doing this is expensive as it requires a skilled human artist to spend a lot of time in an one on one consultation with a witness if it was possible to sit the witness down and talk to a computer to help it construct an image of the suspect this could probably be used in a lot more cases than it can now. Arguably this could potentially give even less potentially contaminated results than using a human too as a human can contaminate the witness testimony subconsciously due to their body language etc a computer couldn't even unintentionally influence the witness towards a desired outcome as it has no desires of it's own let alone non verbal cues by which to let any prejudicial expectations slip though it could do nothing more or less than do it's best to produce a satisfactory match for the image the witness is trying to describe.

  13. There were cases where the GAN that transforms Horses to Zebras cryptographically encoded the original Horse image into the Zebra image, so that the second GAN can then easily reproduce the original image.

  14. Reminds me of when translation software first started becoming popular, the advice was to translate the text then translate it back to see if it still made sense.

  15. The application for MR to CT is radiation therapy. The soft tissue contrast from MRI is often used to determine the target region, but the CT carries the radio-density information needed for determining the radiation therapy dose distribution plan. Instead of taking both MRI and CT and registering them, the synthetic CT from MRI lets you get the job done without the extra time, expense, and radiation dose of the CT.

  16. I can recommend the 2001 SO Picasso version, it's amazing and should have been linked in the description.

  17. So I can see how the loss functions for the individual images would be applied to their respective networks, but how would the loss function between the two horse images be applied? To each network separately? All together as one big system?

  18. That part in the beginning about commissioning artists to paint pictures for the GAN as direct examples… I've long had this idea that Hollywood needs to take digital cameras, and film cameras, and have them shoot at a variety of things in tandem, then run that data through a GAN, so we produce a system to make digital footage look like film footage, and rid ourselves of all the weird digital artifacts we see in films today.

    Mainly I'm just on a campaign to change the way digital photography handles bright coloured lights.

  19. What if you could reduce images into its descriptive parts, then they could be removed from the image….each part can be compressed separately, then you have a higher ratio compression for video….and you can create internet 2.0…lol.

  20. I wonder what happens if you do it with very different images, like horses to oranges back to horses, or Computerfile videos to Numberfile images back to Computerfile videos.

  21. This reminded me of the "game" of translating a text, translating it back and laughing at the differences. Could CycleGAN help improve machine translations?

  22. Here's an idea: Train a GANN to convert to and from Disney art style, to real life. Then Disney can use it to create a live action remake with basically no budget. Instant profit!

  23. For fixing the flickering: DVDs record the things which change between two consecutive images in a movie. Could we train an additional GAN to distinguish between such changes in normal movies (basically DVD data for real movies), perhaps even the original (style-unchanged) movie, and the changes which produce flickers in the style-changed product movies?

  24. I wonder if it is possible to train these systems to undo image processing effects such as a Gaussian blur, and if so how accurate they could be since you could easily create a huge number of paired data points? I know that there are inverse Gaussian blur equations that can be used but these do create a certain amount of error, so it would be interesting to compare the two approaches to see which could produce the most accurate results.

  25. Here's a business oppourtunity: train a GAN to covert my crappy photos into professional photos, and then become a professional photographer.

  26. But isn’t “zebrafying” a horse a lossy process? What I mean is if you convert a brown horse to a zebra and you convert a white horse of the exact same size to a zebra they should produce the same result. But then how will the inverse function know what color to turn it back into?

  27. How do you make sure the pictures you use to train your GAN are actual genuine photos and not computer generated? It's rather improbable you have a million photos of zebras on your hard drive at home and as soon as you use a source like google images, what you effectively get is another ai in the mix.
    And the more pictures of zebras are generated and uploaded to the internet in some way, the worse your data set for training becomes.

    Or if you generalize the problem i guess the the question I'm asking is can humanity as a whole generate enough authentic data to stop generated data from impairing ai training?

  28. How different are horses and zebras? Like Nord-European humans and South-African humans? Or more, or less different?

  29. Oh finally! I've got all these horse pictures left to do and the kindergarten kids are really bad at doing proper shadows, especially around the neck.
    You're a life saver!

  30. Could you use a Gan like this to turn live-action movies into animated movies? Basically the reverse of a disney-remake?

  31. Dr. Pound shoulder shirt tug tally: 6. We need an algorithm to predict if the next video will have increasing or decreasing number of shoulder shirt tug twitches.

  32. I like how he is even optimising the scribble time by painting the generator differently at 5:26 –before in 1:30 was slower.

  33. The real loss needs to calculate how the fake prodused is like intended product. I.e. photo to rembrant painting style, then how well the produced painting is similar to rembrant style..

  34. I guess the results could be improved a bit by adding a few free output neurons that are not part of the zebra image. These could be used by the first GAN to encode the information it destroys by turning a horse to a zebra. This could then be used by the second GAN to reconstruct a better image.
    It should improve the quality of the result because it gives more freedom to the first GAN to modify the image, but not so much that it could just throw a generic zebra.

  35. Dr Mike: "If you've got a lot of pictures of horses and want to turn them into a lot of pictures of zebras, how are you going to do that?…"
    Me: "I've heard enough. Funding approved."

Leave a Reply

Your email address will not be published. Required fields are marked *