😑 Deblurring feels so against the spirit of comeback. Yuck.
March 13, 2026 at 11:40 AM
😑 Deblurring feels so against the spirit of comeback. Yuck.
They basically said they were using a deblurring and sharpening algorithm for shorts, which I agree isn’t technically “GenAI” in the strict sense but I really would like to know what models they’re using for this.
August 26, 2025 at 2:45 AM
They basically said they were using a deblurring and sharpening algorithm for shorts, which I agree isn’t technically “GenAI” in the strict sense but I really would like to know what models they’re using for this.
Cellpose3
github.com/mouseland/ce...
Image Restoration via Denoising (Segmentation loss/Pixel reconstruction loss/Perceptual loss) Deblurring/Deconvolution Upsampling
@computingnature.bsky.social @marius10p.bsky.social @naturemethods.bsky.social 2025
www.nature.com/articles/s41...
github.com/mouseland/ce...
Image Restoration via Denoising (Segmentation loss/Pixel reconstruction loss/Perceptual loss) Deblurring/Deconvolution Upsampling
@computingnature.bsky.social @marius10p.bsky.social @naturemethods.bsky.social 2025
www.nature.com/articles/s41...
February 20, 2025 at 12:15 PM
Cellpose3
github.com/mouseland/ce...
Image Restoration via Denoising (Segmentation loss/Pixel reconstruction loss/Perceptual loss) Deblurring/Deconvolution Upsampling
@computingnature.bsky.social @marius10p.bsky.social @naturemethods.bsky.social 2025
www.nature.com/articles/s41...
github.com/mouseland/ce...
Image Restoration via Denoising (Segmentation loss/Pixel reconstruction loss/Perceptual loss) Deblurring/Deconvolution Upsampling
@computingnature.bsky.social @marius10p.bsky.social @naturemethods.bsky.social 2025
www.nature.com/articles/s41...
Had a play with @elevenlabs.io AI image deblurring. At one level, it is great to be able to save a photo; at another, it really feels like cheating! 🪶
April 15, 2026 at 10:54 AM
Had a play with @elevenlabs.io AI image deblurring. At one level, it is great to be able to save a photo; at another, it really feels like cheating! 🪶
From @laurie-amen.bsky.social: Lenses aren’t always made of glass, sometimes they’re made of dark matter! How can we tell primordial black holes apart from dark black holes using gravitational microlensing. ⚛️ 🔭 ☄️🧪
astrobites.org/2026/04/02/deblurring-dark-compact-objects-with-microlensing/
astrobites.org/2026/04/02/deblurring-dark-compact-objects-with-microlensing/
Deblurring the Dark Compact Object Picture with Microlensing
Lenses aren’t always made of glass, sometimes they’re made of dark matter compact objects! Today’s paper explores if and how we can tell primordial black holes apart from dark black holes using gravit...
astrobites.org
April 6, 2026 at 10:36 PM
From @laurie-amen.bsky.social: Lenses aren’t always made of glass, sometimes they’re made of dark matter! How can we tell primordial black holes apart from dark black holes using gravitational microlensing. ⚛️ 🔭 ☄️🧪
astrobites.org/2026/04/02/deblurring-dark-compact-objects-with-microlensing/
astrobites.org/2026/04/02/deblurring-dark-compact-objects-with-microlensing/
Sihan Wang, Jinshu Huang, Haibin Su, Yunhua Xue: Frequency-Decoupled Diffusion Guidance for Non-Blind Image Deblurring https://arxiv.org/abs/2610.06221 https://arxiv.org/pdf/2610.06221 https://arxiv.org/html/2610.06221
October 6, 2026 at 6:43 AM
Sihan Wang, Jinshu Huang, Haibin Su, Yunhua Xue: Frequency-Decoupled Diffusion Guidance for Non-Blind Image Deblurring https://arxiv.org/abs/2610.06221 https://arxiv.org/pdf/2610.06221 https://arxiv.org/html/2610.06221
Another super paper came out this week from Amit Kohli and Laura Waller on Ring deconvolution microscopy. This one is fairly technical, but essentially it's a computational approach that lets you do deconvolution when there is a spatially varying PSF. www.nature.com/articles/s41...
Ring deconvolution microscopy: exploiting symmetry for efficient spatially varying aberration correction - Nature Methods
Ring deconvolution microscopy leverages symmetry to provide fast and straightforward spatially varying deblurring and offers improved speed and image quality compared to standard approaches across div...
www.nature.com
May 2, 2025 at 1:23 PM
Another super paper came out this week from Amit Kohli and Laura Waller on Ring deconvolution microscopy. This one is fairly technical, but essentially it's a computational approach that lets you do deconvolution when there is a spatially varying PSF. www.nature.com/articles/s41...
The website might have also used ai which is cringe. Yet another L, im way prettier than this. Anyways, i was kinda scared of deblurring technology until now. Deblurring is kinda spooky if i blur something private, but i dont think i have much to worry about.
May 3, 2026 at 5:07 AM
The website might have also used ai which is cringe. Yet another L, im way prettier than this. Anyways, i was kinda scared of deblurring technology until now. Deblurring is kinda spooky if i blur something private, but i dont think i have much to worry about.
Almost couldn’t wait until #FluorescenceFriday to post this one 😄🧪
Teeny tiny Golgi wiggles (if you squint 🧐)
Excited to sink my teeth into more brightfield microscopy!! Good time resolution + deblurring = 1 very happy PhD student 🥳🧫🔬
Teeny tiny Golgi wiggles (if you squint 🧐)
Excited to sink my teeth into more brightfield microscopy!! Good time resolution + deblurring = 1 very happy PhD student 🥳🧫🔬
November 15, 2024 at 6:07 AM
Almost couldn’t wait until #FluorescenceFriday to post this one 😄🧪
Teeny tiny Golgi wiggles (if you squint 🧐)
Excited to sink my teeth into more brightfield microscopy!! Good time resolution + deblurring = 1 very happy PhD student 🥳🧫🔬
Teeny tiny Golgi wiggles (if you squint 🧐)
Excited to sink my teeth into more brightfield microscopy!! Good time resolution + deblurring = 1 very happy PhD student 🥳🧫🔬
They have some futuristic deblurring tech the public isn't aware of... Next time, I'm just cropping the bottom part off!
November 26, 2024 at 5:21 PM
They have some futuristic deblurring tech the public isn't aware of... Next time, I'm just cropping the bottom part off!
YT said they'd have an opt-out "soon" - in August. Like, shouldn't you have built an opt-out into it from the start?
Oh, right! Google doesn't believe in consent.
www.nbcnews.com/tech/tech-ne...
Oh, right! Google doesn't believe in consent.
www.nbcnews.com/tech/tech-ne...
YouTube addresses creators' concerns that it secretly used AI to edit some videos
"Creators, we’ve heard your feedback on YouTube’s deblurring and denoising Shorts," Rene Ritchie, YouTube’s creator liaison, wrote on X.
www.nbcnews.com
October 14, 2025 at 11:57 AM
YT said they'd have an opt-out "soon" - in August. Like, shouldn't you have built an opt-out into it from the start?
Oh, right! Google doesn't believe in consent.
www.nbcnews.com/tech/tech-ne...
Oh, right! Google doesn't believe in consent.
www.nbcnews.com/tech/tech-ne...
Spatially variant deconvolution by Amit Kohli et al:
www.nature.com/articles/s41...
Can be used advantageously with light sheet fluorescence microscopy.
www.nature.com/articles/s41...
Can be used advantageously with light sheet fluorescence microscopy.
Ring deconvolution microscopy: exploiting symmetry for efficient spatially varying aberration correction - Nature Methods
Ring deconvolution microscopy leverages symmetry to provide fast and straightforward spatially varying deblurring and offers improved speed and image quality compared to standard approaches across div...
www.nature.com
April 29, 2025 at 10:51 PM
Spatially variant deconvolution by Amit Kohli et al:
www.nature.com/articles/s41...
Can be used advantageously with light sheet fluorescence microscopy.
www.nature.com/articles/s41...
Can be used advantageously with light sheet fluorescence microscopy.
specially it lets me use a deblurring algorithm that uses non-generative techniques in its ai which is uh, something I value
September 2, 2025 at 3:40 AM
specially it lets me use a deblurring algorithm that uses non-generative techniques in its ai which is uh, something I value
Junseong Shin, Hyeonsu Jo, Daehyun Kim, Tae Hyun Kim: PickMoment: Continuous-Time Single-Image-to-Video via Learning Deblurring and Blur-to-Video https://arxiv.org/abs/2610.01279 https://arxiv.org/pdf/2610.01279 https://arxiv.org/html/2610.01279
October 2, 2026 at 6:41 AM
Junseong Shin, Hyeonsu Jo, Daehyun Kim, Tae Hyun Kim: PickMoment: Continuous-Time Single-Image-to-Video via Learning Deblurring and Blur-to-Video https://arxiv.org/abs/2610.01279 https://arxiv.org/pdf/2610.01279 https://arxiv.org/html/2610.01279
The thing about deblurring composite video is that the blur is important for 3D games. Harsh edges on your pixels isn't great in a 3D game
November 5, 2024 at 10:18 AM
The thing about deblurring composite video is that the blur is important for 3D games. Harsh edges on your pixels isn't great in a 3D game
🎉 New paper: SSMB
A self-supervised keypoint detector that works directly on motion-blurred images
No deblurring preprocessing, no handcrafted detector, no external pseudo-labels
📄 Paper: arxiv.org/abs/2608.27181
#ComputerVision #KeypointDetection #SelfSupervisedLearning #ImageMatching
A self-supervised keypoint detector that works directly on motion-blurred images
No deblurring preprocessing, no handcrafted detector, no external pseudo-labels
📄 Paper: arxiv.org/abs/2608.27181
#ComputerVision #KeypointDetection #SelfSupervisedLearning #ImageMatching
August 31, 2026 at 4:27 AM
🎉 New paper: SSMB
A self-supervised keypoint detector that works directly on motion-blurred images
No deblurring preprocessing, no handcrafted detector, no external pseudo-labels
📄 Paper: arxiv.org/abs/2608.27181
#ComputerVision #KeypointDetection #SelfSupervisedLearning #ImageMatching
A self-supervised keypoint detector that works directly on motion-blurred images
No deblurring preprocessing, no handcrafted detector, no external pseudo-labels
📄 Paper: arxiv.org/abs/2608.27181
#ComputerVision #KeypointDetection #SelfSupervisedLearning #ImageMatching
New paper: "Deblurring fission fragment mass distributions", link: https://arxiv.org/abs/2505.01294
December 2, 2025 at 8:15 PM
New paper: "Deblurring fission fragment mass distributions", link: https://arxiv.org/abs/2505.01294
Resources for image and video deblurring
April 18, 2026 at 12:25 PM
Resources for image and video deblurring
Diffusion models are sick, right? They're a million times more interesting than LLMs! I'm trying to understand renormalization in quantum field theory and I keep getting similar vibes. Like sending the momentum cutoff to infinity is analogous to deblurring. Wish I understood both better.
April 17, 2025 at 2:00 AM
Diffusion models are sick, right? They're a million times more interesting than LLMs! I'm trying to understand renormalization in quantum field theory and I keep getting similar vibes. Like sending the momentum cutoff to infinity is analogous to deblurring. Wish I understood both better.
RefDeblur: Blind Motion Deblurring with Self-Generated Reference Image
Insoo Kim, Geonseok Seo, Hyong-Euk Lee, Jinwoo Shin
Action editor: Robert Legenstein
https://openreview.net/forum?id=Nyewu7xztw
#blurring #deblurring #blur
Insoo Kim, Geonseok Seo, Hyong-Euk Lee, Jinwoo Shin
Action editor: Robert Legenstein
https://openreview.net/forum?id=Nyewu7xztw
#blurring #deblurring #blur
November 16, 2025 at 9:18 AM
RefDeblur: Blind Motion Deblurring with Self-Generated Reference Image
Insoo Kim, Geonseok Seo, Hyong-Euk Lee, Jinwoo Shin
Action editor: Robert Legenstein
https://openreview.net/forum?id=Nyewu7xztw
#blurring #deblurring #blur
Insoo Kim, Geonseok Seo, Hyong-Euk Lee, Jinwoo Shin
Action editor: Robert Legenstein
https://openreview.net/forum?id=Nyewu7xztw
#blurring #deblurring #blur
New paper: "Deblurring fission fragment mass distributions", link: https://arxiv.org/abs/2505.01294
May 5, 2025 at 8:15 PM
New paper: "Deblurring fission fragment mass distributions", link: https://arxiv.org/abs/2505.01294
When you teach a mini-course, and for their final project, your students work on deblurring the image of a 'mate'... you don't even need ChatGPT to guess where they’re from. 🤣🤣🤣🤣
December 18, 2024 at 2:06 PM
When you teach a mini-course, and for their final project, your students work on deblurring the image of a 'mate'... you don't even need ChatGPT to guess where they’re from. 🤣🤣🤣🤣