Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

Can we enable free-viewpoint renderings of scenes under various lighting conditions? Yes, we can! And just from SINGLE videos. Here are 🌃 night simulations from input videos with hard shadows.

jbhuang0604's profile picture

Jia-Bin Huang

79,561 subscribers

26,297 görüntüleme • 1 yıl önce •via X (Twitter)

Bilim & TeknolojiEğitimSanat
LIVE

Anya Rossi• Live Now

Private livecam show

6 Yorum

Jia-Bin Huang profil fotoğrafı
Jia-Bin Huang1 yıl önce

To reconstruct and re-render the scene under different lighting, we need Inverse Rendering. We need to infer shape, albedo, visibility, and sun and sky illumination from videos. This differs from existing NeRF/3DGS, where the lighting was baked into the models.

Jia-Bin Huang profil fotoğrafı
Jia-Bin Huang1 yıl önce

How? 🤔 Let's integrate physics and learning! 🤝 Physically based rendering helps explain away shadow volume estimation. Learning-based priors: Help avoid degenerated inverse rendering solutions.

Jia-Bin Huang profil fotoğrafı
Jia-Bin Huang1 yıl önce

Inverse rendering enables many exciting new applications beyond view synthesis. In the paper, we showcase 🌉 nighttime simulation 📽️ object insertion 💡 changing sun directions

Jia-Bin Huang profil fotoğrafı
Jia-Bin Huang1 yıl önce

💻 Code is available! Have fun! We will be presenting this work at @3DVconf!

Jia-Bin Huang profil fotoğrafı
Jia-Bin Huang1 yıl önce

Check out the project page for more results and comparisons. A fun collaboration between 🌽 and 🐢!

darks profil fotoğrafı
darks1 yıl önce

Very interesting. What paper is this from???

Benzer Videolar

oh shit, here we go again! Animate Anyone 2 can replace people in videos with anyone from a single reference image 🔥
0:51

oh shit, here we go again! Animate Anyone 2 can replace people in videos with anyone from a single reference image 🔥

Dreaming Tulpa 🥓👑

253,017 görüntüleme • 1 yıl önce

(1/2) LightIt: Illumination Modeling and Control for Diffusion Models! #CVPR2024 We facilitate lighting control for novel image generation from text prompts. We can also edit lighting for a given input image. Video: Project:
0:31

(1/2) LightIt: Illumination Modeling and Control for Diffusion Models! #CVPR2024 We facilitate lighting control for novel image generation from text prompts. We can also edit lighting for a given input image. Video: Project:

Matthias Niessner

19,849 görüntüleme • 2 yıl önce

“If you are in support of Pal€stine🇵🇸, then you are breaching conditions & you can be arrested.” “& if we are in support of 1$rael, we can stay?” “Yes, Yes”
0:51

“If you are in support of Pal€stine🇵🇸, then you are breaching conditions & you can be arrested.” “& if we are in support of 1$rael, we can stay?” “Yes, Yes”

Censored Humans

709,998 görüntüleme • 6 ay önce

LEAKED DOLLHOUSE D!DDY HOME VIDEOS ARE INSANE & DISGUSTING❗ D!ddy has videos leak from behind the scenes and from television and they are DAMNING now that we know what we know! CANST
34:34

Sensitive content

LEAKED DOLLHOUSE D!DDY HOME VIDEOS ARE INSANE & DISGUSTING❗ D!ddy has videos leak from behind the scenes and from television and they are DAMNING now that we know what we know! CANST

Jasper Truth 🇺🇸 🇨🇦

1,355,064 görüntüleme • 1 yıl önce

Can we make full-length videos with the help of AI? We explored this with various AI tools and here's what we came up with. Sound on! (Full workflow & secret tool below)
0:38

Can we make full-length videos with the help of AI? We explored this with various AI tools and here's what we came up with. Sound on! (Full workflow & secret tool below)

The Deep View

247,265 görüntüleme • 1 yıl önce

AI Bedtimestory You can soon create personal audio/image stories like this from a single input ⚡️ It's been more than 2 years since we launched Bedtimestory. Some +100.000 hyper personal stories later, and here we are. But it's just not limited to bedtimestories... 🧵
6:05

AI Bedtimestory You can soon create personal audio/image stories like this from a single input ⚡️ It's been more than 2 years since we launched Bedtimestory. Some +100.000 hyper personal stories later, and here we are. But it's just not limited to bedtimestories... 🧵

Linus ✦ Ekenstam

14,029 görüntüleme • 1 yıl önce

I just recreated iconic scenes from Game of Thrones using Veo 2 from Google DeepMind! How many can you recognize? We can finally make the ending we deserve.
1:46

I just recreated iconic scenes from Game of Thrones using Veo 2 from Google DeepMind! How many can you recognize? We can finally make the ending we deserve.

Ammaar Reshi

260,847 görüntüleme • 1 yıl önce

Meet Gemini Omni, our new model that can create anything from any input, starting with video. With Gemini Omni, you can combine images, videos and text as inputs and generate high-quality videos grounded in Gemini's real-world knowledge.
0:50

Meet Gemini Omni, our new model that can create anything from any input, starting with video. With Gemini Omni, you can combine images, videos and text as inputs and generate high-quality videos grounded in Gemini's real-world knowledge.

Google Gemini

32,779,480 görüntüleme • 1 ay önce

Meet Gemini Omni, our new model that can create anything from any input, starting with video. With Gemini Omni, you can combine images, videos and text as inputs and generate high-quality videos grounded in Gemini's real-world knowledge. #GoogleIO
0:50

Meet Gemini Omni, our new model that can create anything from any input, starting with video. With Gemini Omni, you can combine images, videos and text as inputs and generate high-quality videos grounded in Gemini's real-world knowledge. #GoogleIO

Google Gemini

88,780 görüntüleme • 1 ay önce

Luma's start and end keyframes are a game changer. With a sequence of keyframes from the original film, we can seamlessly remaster stop motion classics like "Jason and the Argonauts" as modern single-take action scenes.
1:01

Luma's start and end keyframes are a game changer. With a sequence of keyframes from the original film, we can seamlessly remaster stop motion classics like "Jason and the Argonauts" as modern single-take action scenes.

Jonathan Fly 👾

5,414,758 görüntüleme • 2 yıl önce

AI is a massive unlock for medicine because once you can simulate cells/human biology you can fast track cures for almost all diseases. Precision medicine becomes real. But if we can simulate cells then we are not far off from simulating the world Are we going to be running simulations within simulation? How many simulations deep are we?
1:48

AI is a massive unlock for medicine because once you can simulate cells/human biology you can fast track cures for almost all diseases. Precision medicine becomes real. But if we can simulate cells then we are not far off from simulating the world Are we going to be running simulations within simulation? How many simulations deep are we?

Andrew Kang

54,207 görüntüleme • 1 yıl önce

Kling 2.6 has arrived. Create AI videos from text and videos with native audio, SFX and music at fantastic quality. When combined with Kling O1 and Kling Omni, you can get multiple angles of a scene all from a singular video input. Here's my tutorial!
3:49

Kling 2.6 has arrived. Create AI videos from text and videos with native audio, SFX and music at fantastic quality. When combined with Kling O1 and Kling Omni, you can get multiple angles of a scene all from a singular video input. Here's my tutorial!

Jerrod Lew

12,139 görüntüleme • 7 ay önce

When recognize that we are the ones creating the reaction, we can begin to free our true consciousness. We are the creators of our life, and can create the quality of life we want from anything life offers us. #DateWithDestiny2025
0:58

When recognize that we are the ones creating the reaction, we can begin to free our true consciousness. We are the creators of our life, and can create the quality of life we want from anything life offers us. #DateWithDestiny2025

Tony Robbins

60,873 görüntüleme • 7 ay önce

GEN-1 still works with lights off, and generalizes under harsh lighting conditions. The model uses raw video pixels to make decisions, so strong lighting changes can drastically alter its input distribution. Yet performance still holds. Why? GEN-1 was pre-trained on a massive, diverse dataset of different lighting conditions—everywhere from outdoor farms, to warehouses, from grocery stores, to dimly lit homes—it's already seen it all, and transfers this knowledge to new tasks. This is a glimpse of what we call Mastery, and is part of the reason these models can cross a new performance threshold. Read more about it in our blog post in the comments below 👇
0:39

GEN-1 still works with lights off, and generalizes under harsh lighting conditions. The model uses raw video pixels to make decisions, so strong lighting changes can drastically alter its input distribution. Yet performance still holds. Why? GEN-1 was pre-trained on a massive, diverse dataset of different lighting conditions—everywhere from outdoor farms, to warehouses, from grocery stores, to dimly lit homes—it's already seen it all, and transfers this knowledge to new tasks. This is a glimpse of what we call Mastery, and is part of the reason these models can cross a new performance threshold. Read more about it in our blog post in the comments below 👇

Generalist

18,101 görüntüleme • 2 ay önce

Can someone go to Thembisa hospital and get some videos of the conditions there. So we can do a split screen of how Hangwani lives and how his victims suffer.
1:15

Can someone go to Thembisa hospital and get some videos of the conditions there. So we can do a split screen of how Hangwani lives and how his victims suffer.

Africa Research Desk

779,762 görüntüleme • 9 ay önce

We are so cooked. AI videos are becoming indistinguishable from reality and this is just the beginning
0:58

We are so cooked. AI videos are becoming indistinguishable from reality and this is just the beginning

Ian Miles Cheong

42,365 görüntüleme • 1 yıl önce

“Dignity”? How about SOVEREIGNTY. We are so done with the lies. We are so done with the fraud. We are so done with the insanity and the excuses. My 20/20 Sovereignty Act: 20 years of ZERO immigration, 20 million deportations. Yes, we can. Elect patriots—we can just do things.
5:00

“Dignity”? How about SOVEREIGNTY. We are so done with the lies. We are so done with the fraud. We are so done with the insanity and the excuses. My 20/20 Sovereignty Act: 20 years of ZERO immigration, 20 million deportations. Yes, we can. Elect patriots—we can just do things.

John Strand

60,690 görüntüleme • 3 ay önce

[SIGGRAPH '25] EVA: Expressive Virtual Avatars from Multi-view Videos Contributions: 1. We introduce EVA, a novel method enabling full-body control with real-time, photo-realistic renderings, robustly handling loose clothing dynamics and various facial expressions. 2. We develop an expressive deformable template that generates a deformable human template mesh and employs a multi-stage tracking algorithm to faithfully capture facial expressions, body motions, and non-rigid deformations from multi-view videos. 3. We propose a disentangled 3D Gaussian appearance module that models the body and face independently, ensuring separated control and high-quality renderings.
1:22

[SIGGRAPH '25] EVA: Expressive Virtual Avatars from Multi-view Videos Contributions: 1. We introduce EVA, a novel method enabling full-body control with real-time, photo-realistic renderings, robustly handling loose clothing dynamics and various facial expressions. 2. We develop an expressive deformable template that generates a deformable human template mesh and employs a multi-stage tracking algorithm to faithfully capture facial expressions, body motions, and non-rigid deformations from multi-view videos. 3. We propose a disentangled 3D Gaussian appearance module that models the body and face independently, ensuring separated control and high-quality renderings.

MrNeRF

18,407 görüntüleme • 1 yıl önce