[HN Gopher] Deblur-GS: 3D Gaussian splatting from camera motion ...
___________________________________________________________________
Deblur-GS: 3D Gaussian splatting from camera motion blurred images
Author : smusamashah
Score : 91 points
Date : 2024-05-13 17:16 UTC (5 hours ago)
(HTM) web link (chaphlagical.icu)
(TXT) w3m dump (chaphlagical.icu)
| nathancahill wrote:
| I know it's a meme at this point but this is real life "Enhance
| please". Incredibly impressive what we're able to do to
| reconstruct missing data.
| zevv wrote:
| My friend doesn't quite grasp this yet, can someone explain? Is
| the reconstructed detail all "real" and extracted from the
| blurred input, or is there some model at work here, filling in
| the image with plausible details, but basically making up stuff
| that was not really there to start with?
| karmakaze wrote:
| That's accurate. What's worth nothing though is that everything
| we 'see' with our own eyes is constructed from sampling our
| environment. The image we construct is what we expected to see
| given the sample data. This is one reason why eyewitness
| testimony can be vivid and false without any foul play.
| barrysteve wrote:
| Both. The paper mentions using a deblurrer and novel view
| synthesis model(ExBluRF).
| tomp wrote:
| I skimmed the Overview and am not an expert.
|
| It seems to me they don't use any ML at all. They use
| backpropagation to jointly optimise the entire physics/motion
| model, which models camera motion and the generated blurry
| images (they generate multiple images for each camera frame
| along the path of motion of the camera, and then merge them,
| simulating motion blur)
| dheera wrote:
| It is ML in the sense of optimizing a nonconvex loss function
| over a dataset. It is not a fancy diffusion model or even a
| generative model, but it is no less a machine learning
| problem.
| tomp wrote:
| "Not ML" as in "not learning from data to apply in new
| situations" but rather they do "mathematical optimisation".
|
| The data they optimise over is just the images of the
| current camera trajectory (as far as I understand)
| chpatrick wrote:
| Gaussian blur is a reversible operation, but in practice it's
| not possible on still images. With multiple pictures you might
| have enough information.
| peppertree wrote:
| No it does not "make up things" using generative AI. Current GS
| implementations assume camera poses are static. This paper
| assigns a linear motion trajectory to camera during training.
| creativeSlumber wrote:
| So can it handle when both camera and multiple objects in
| scene are moving in different trajectories?
| dheera wrote:
| Not with traditional 3D Gaussian splatting, but it is
| potentially possible to separate the time axis and do a 4D
| Gaussian splatting with some regularization to accommodate
| dynamic scenes.
|
| Here's some early work in this area which seems promising:
| https://guanjunwu.github.io/4dgs/
| emilk wrote:
| Very cool! A next step could be to model a rolling-shutter
| tomaskafka wrote:
| Absolutely impressive - seems on par with what's happening in our
| eyes and brain. If this becomes realtime, we could turn the noisy
| low fps image from cameras on AR headsets in dark environments
| into smooth bright image.
| borgchick wrote:
| finally, all the UFO videos can be clear!
| bee_rider wrote:
| The aliens are actually pan-dimensional light beings. That is
| why they are afraid of high quality cameras, if they get caught
| in a photo they are stuck here forever. Running this algorithm
| on pictures of UFOs is actually an intergalactic warcrime.
| germinator wrote:
| I really want to be impressed, but I've been reading papers about
| breakthroughs in deblurring and upscaling for two decades now,
| and the state of the art in commercial and open-source tools is
| still pretty underwhelming. Chances are, if you have a low-res
| keepsake photo, or take a blurry nature shot, you're gonna be
| stuck with that.
|
| Video, where the result needs to be temporally coherent and make
| sense in 3D, can't be the easier one.
| tomp wrote:
| this work won't solve that. it requires a video (sequence of
| images)
| IshKebab wrote:
| > Video, where the result needs to be temporally coherent and
| make sense in 3D, can't be the easier one.
|
| Why not? Video is a much more tractable problem because you
| have much more information to go on.
| adkaplan wrote:
| https://web.archive.org/web/20240511220923/https://chaphlagi...
|
| Down for me, archive above.
| smusamashah wrote:
| The reconstruction looks even better than ground truth images in
| their examples.
___________________________________________________________________
(page generated 2024-05-13 23:00 UTC)