Computer: Compression
Boob friendly video compression layer   (0)  [vote for, against]
Recognize faces and improve compression with them

One method for lossy video compression is to take two frames that are duplicates or near duplicates and merge them, saving only one frame. For example if a person is talking in front of a wall, there's no need to save every single frame of that wall. You can take one picture of it and just use that until the lighting changes substantially or the camera moves. That's all well and good.

But for humans some parts of an image are much more important than others. A person's eyes may be in nearly the exact same position, but a couple of millimeters of movement there is important to communication. You can tell whether someone is looking at your eyes or at your nose from across the room.

Modern compression algorithms tend to throw out the baby with the bathwater. They'll happily compress the wall with the same vigor as the unmoving sword in a swordfight. A tiny twitch of the sword is important, but a slight shadow on the wall is not.

So the idea is to add to video compression a recognition of the most important parts of a video. All human or animal movements are important no matter how small. But set pieces are much less so. Face recognition is already done by cameras for the purpose of focus and so forth. This video compression library would add thousands of probably-important general shapes to mark for less lossy compression, allowing higher effective fidelity at lower file sizes.
-- Voice, Dec 21 2020

https://en.wikipedi...Jeremy_Hillary_Boob [pertinax, Dec 22 2020]

I was told there would be boobs.
-- FlyingToaster, Dec 21 2020


Among other shapes, yes. It's not just eyes and swords one wants not to be obscured.
-- Voice, Dec 21 2020


Who cares about swords or eyes
-- pocmloc, Dec 21 2020


//A tiny twitch of the sword is important, but a slight shadow on the wall is not.//

I'm not sure I like this anti-wall agenda. As a wall enthusiast, I'm often dismayed when compressed videos remove the complex surface structure and subtle lighting variations of a really first-rate wall by displaying it as a handful of massive pixels. I mean, people have only been waving swords around for a few thousand years, walls go back much further. No swords at Gobekli tepe, walls galore though. What's a pyramid if not a complex 3D supporting structure for internal passageways and rooms?.. a big pointy wall.
-- bs0u0155, Dec 21 2020


From the title, we thought this might be some sort of thin, close fitting, elasticated garment for females, or a rather tawdry exploitative video employing such garments.
-- 8th of 7, Dec 21 2020


I would think that sword enthusiasts might be as interested in the minute fluctuations of light and shadow playing across the wall - they might even give a more nuanced sense of the sword's movement than the straight on view of the sword.
-- pocmloc, Dec 21 2020


Toaster, You turned left instead of right...you want the Half- boobery. It's down the hall...waaaay down!
-- blissmiss, Dec 21 2020


One thing about this Human Salient Compression (HSC) [Voice] came up with is that it really has varied commercial applicability. I think I figured out how to do it technologically (too long for here) and noticed that if Disney compresses their entire 100,000 movie/show back catalog that is like 5% of Netflix compressing their entire back catalog. So Disney gets to go "all high quality look" on everything for 20 times less money.

Youtube: sadly, no originals to recompress. (?)

Porn: Easy to recompress sort of uninspiring mass produced porn, but originals on great amateur stuff are not around to recompress.

Aside: If you can compress on salience, you've identified salience, and can turn salience up. That spices the videos. Unknown if it makes them better, it does make people's brains react to them more.
-- beanangel, Dec 22 2020


One way to do this would be to use eye-tracking devices on all the people watching a test screening of this movie. This would record the precise parts of the image in each scene to which people gave most attention. Then, the compression algorithm would use this data, applying heavy compression to those parts of the scene to which people gave little attention, and very light compression to the bits which people were looking at (which are presumably those elements most relevant to the plot, like the sword you mention).
-- hippo, Dec 22 2020


//I was told there would be boobs//

Well, if you include people bearing some passing resemblance to Jeremy Hillary Boob, Ph.D., there are.
-- pertinax, Dec 22 2020



random, halfbakery