LogoThread Easy
  • Explore
  • Thread Compose
LogoThread Easy

Your All-in-One Twitter Thread Companion

© 2025 Thread Easy All Rights Reserved.

Explore

Newest first — browse tweet threads

Keep on to blur preview images; turn off to show them clearly

Relightable Holoported Characters: Capturing and Relighting Dynamic Human Performance from Sparse Views

Contributions:
• Relightable Holoported Characters: The first method for photorealistic human performance relighting of unseen motions from sparse driving cameras at inference.

• RelightNet and Physics-informed Feature Maps: A transformer-based network that efficiently computes the rendering equation in a single feed-forward pass. It models light transport as cross-attention between environment maps and local physics-informed features, encoding geometry, albedo, shading, and viewing direction.

• Dataset and Capture Strategy: A data capture strategy and a large-scale, multi-view lightstage dataset for dynamic full-body human relighting. It features five subjects under more than 1000 natural lighting conditions, which we plan to publicly release as a benchmark.

Relightable Holoported Characters: Capturing and Relighting Dynamic Human Performance from Sparse Views Contributions: • Relightable Holoported Characters: The first method for photorealistic human performance relighting of unseen motions from sparse driving cameras at inference. • RelightNet and Physics-informed Feature Maps: A transformer-based network that efficiently computes the rendering equation in a single feed-forward pass. It models light transport as cross-attention between environment maps and local physics-informed features, encoding geometry, albedo, shading, and viewing direction. • Dataset and Capture Strategy: A data capture strategy and a large-scale, multi-view lightstage dataset for dynamic full-body human relighting. It features five subjects under more than 1000 natural lighting conditions, which we plan to publicly release as a benchmark.

Paper: https://t.co/NiIc6b1VcI Project: https://t.co/gghifts0X7

avatar for MrNeRF
MrNeRF
Tue Dec 02 07:49:06
我主要是没有能力做一个专业博主, 我一直把宝玉老师当作榜样, 但是没有学成。  

新人起号,开始阶段, 最重要的是能坚持一下, 能坚持下去的最重要的,你要喜欢这个才可以。 

先不要给自己框住, 想到什么就发什么, AI 泛烂的时候, 真正的活人感反而是稀缺的。

我主要是没有能力做一个专业博主, 我一直把宝玉老师当作榜样, 但是没有学成。 新人起号,开始阶段, 最重要的是能坚持一下, 能坚持下去的最重要的,你要喜欢这个才可以。 先不要给自己框住, 想到什么就发什么, AI 泛烂的时候, 真正的活人感反而是稀缺的。

独立开发者 | 个人IP教练 | 帮助新手在X上完成早期成长| 公众号:PandaTalk8

avatar for Mr Panda
Mr Panda
Tue Dec 02 07:48:05
知名扩展程序 #WeTab 和 Infinity V + 等被发现恶意劫持用户并收集用户方方面面的信息,监控用户访问的任意网站甚至可以发起中间人攻击进行劫持。WeTab 开发商是重庆哨兵电子商务科技有限公司,目前尚不清楚是这些扩展程序开发者主动投毒还是公司被收购后进行的投毒,但考虑到投毒从 2018 年就开始,大概率不是收购造成的。

知名扩展程序 #WeTab 和 Infinity V + 等被发现恶意劫持用户并收集用户方方面面的信息,监控用户访问的任意网站甚至可以发起中间人攻击进行劫持。WeTab 开发商是重庆哨兵电子商务科技有限公司,目前尚不清楚是这些扩展程序开发者主动投毒还是公司被收购后进行的投毒,但考虑到投毒从 2018 年就开始,大概率不是收购造成的。

查看全文:https://t.co/D9QECNFTTr

avatar for 蓝点网
蓝点网
Tue Dec 02 07:46:26
RT @SakanaAILabs: 【経営企画メンバー募集!】

Corporate Development
https://t.co/gCj68Zr9Mh

Sakana AIは、研究開発と事業開発の両輪で成長をとげ、日本発でありながら既にグローバルAI企業や投資家からも注目…

RT @SakanaAILabs: 【経営企画メンバー募集!】 Corporate Development https://t.co/gCj68Zr9Mh Sakana AIは、研究開発と事業開発の両輪で成長をとげ、日本発でありながら既にグローバルAI企業や投資家からも注目…

Building @SakanaAILabs 🧠

avatar for hardmaru
hardmaru
Tue Dec 02 07:45:08
No video game manages to beat the open world rpg game of business. It’s just so fun and addicting. You can pick any lane (music, engineering, fashion) and compete till you’re top 1% while meeting the best along the way.

No video game manages to beat the open world rpg game of business. It’s just so fun and addicting. You can pick any lane (music, engineering, fashion) and compete till you’re top 1% while meeting the best along the way.

Founder: @mixpanel Pizzatarian, programmer, music maker

avatar for Suhail
Suhail
Tue Dec 02 07:44:14
EGG-Fusion: Efficient 3D Reconstruction with Geometry-Aware Gaussian Surfel on the Fly

Contributions:
• We propose EGG-Fusion, a Gaussian surfel-based real-time 3D reconstruction system for high-quality scene modeling in real-world environments. We release the source code to facilitate research reproducibility and community advancement.

• We introduce an information-filter-based surfel fusion method that achieves stable, high-precision reconstruction by dynamically updating geometric information matrices, while significantly accelerating optimization convergence.

• Extensive experimental results demonstrate that the proposed method outperforms current GS-based SOTA methods in tracking accuracy, reconstruction quality, and rendering quality, while running in real-time at 24 FPS.

EGG-Fusion: Efficient 3D Reconstruction with Geometry-Aware Gaussian Surfel on the Fly Contributions: • We propose EGG-Fusion, a Gaussian surfel-based real-time 3D reconstruction system for high-quality scene modeling in real-world environments. We release the source code to facilitate research reproducibility and community advancement. • We introduce an information-filter-based surfel fusion method that achieves stable, high-precision reconstruction by dynamically updating geometric information matrices, while significantly accelerating optimization convergence. • Extensive experimental results demonstrate that the proposed method outperforms current GS-based SOTA methods in tracking accuracy, reconstruction quality, and rendering quality, while running in real-time at 24 FPS.

Paper: https://t.co/bQHSDVBMBI Project: https://t.co/NoUlG0z8rD

avatar for MrNeRF
MrNeRF
Tue Dec 02 07:41:14
  • Previous
  • 1
  • More pages
  • 1829
  • 1830
  • 1831
  • More pages
  • 5634
  • Next