Tencent HY (@TencentHunyuan)

Tencent의 HY 3D 3.1이 글로벌 플랫폼에 배포되었습니다. 이번 업데이트는 텍스처 충실도와 지오메트리 정밀도를 크게 향상시키고 최대 8개의 입력 뷰를 지원해 재구성 정확도와 조각 수준의 디테일을 구현합니다. 신규 크리에이터에게는 20회 무료 이용권이 제공됩니다.

https://x.com/TencentHunyuan/status/2016449283428659599

#tencent #hy3d #3dreconstruction #texture

Tencent HY (@TencentHunyuan) on X

Tencent HY 3D 3.1 is now available on our global platform! This update delivers a massive leap in texture fidelity and geometry precision. It also supports up to 8 input views for ultimate reconstruction accuracy and sculpt-level detail. 💎 New creators can now access 20 free

X (formerly Twitter)

Tencent HY (@TencentHunyuan)

Tencent의 HY 3D 3.1이 글로벌 플랫폼에 공개되었습니다. 이번 업데이트는 텍스처 충실도와 지오메트리 정밀도를 크게 향상시키고, 최대 8개의 입력 뷰를 지원해 재구성 정확도와 조각 수준 디테일을 제공하며, 신규 크리에이터 대상 20개의 무료 제공 등 접근성 개선을 포함합니다.

https://x.com/TencentHunyuan/status/2016449283428659599

#tencent #3dreconstruction #multiview #texture

Tencent HY (@TencentHunyuan) on X

Tencent HY 3D 3.1 is now available on our global platform! This update delivers a massive leap in texture fidelity and geometry precision. It also supports up to 8 input views for ultimate reconstruction accuracy and sculpt-level detail. 💎 New creators can now access 20 free

X (formerly Twitter)

Bài báo mới D4RT giới thiệu mô hình transformer duy nhất, cho phép suy đoán độ sâu, tương quan thời gian‑không gian và tham số camera từ một video duy nhất. Cơ chế truy vấn mới giảm tải tính toán, đồng thời đạt hiệu suất mới trong tái tạo 4D, vượt trội so với các phương pháp trước. #AI #ComputerVision #DeepLearning #3DReconstruction #TríTuệNhânTạo #ThịGiácMáy #TáiTạo3D

https://deepmind.google/blog/d4rt-teaching-ai-to-see-the-world-in-four-dimensions/

D4RT: Unified, Fast 4D Scene Reconstruction & Tracking

Meet D4RT, a unified AI model for 4D scene reconstruction and tracking.

Google DeepMind
Europeana’s EUreka3D is reconstructing Europe’s cultural legacy in 3D and making it interoperable across institutions.
By merging archives, models, and metadata, it’s turning scattered memory into a shared spatial archive, usable for research, education, and public storytelling.
https://pro.europeana.eu/project/eureka3d-european-union-s-rekonstructed-content-in-3d
#DigitalHeritage #3DReconstruction #CulturalMemory
EUreka3D - European Union's REKonstructed content in 3D | Europeana PRO

EUreka3D built the capacity of small cultural heritage institutions in digital transformation, particularly on issues related to 3D digitisation.

Europeana PRO
New AI model turns photos into explorable 3D worlds, with caveats

Openly available AI tool creates steerable 3D-like video, but requires serious GPU muscle.

Ars Technica

The Dispersed Chinese Art Digitization Project bridges centuries and continents through digital reconstructions of major historical sites and artifacts – from the Zhihua Temple to the Six Horses of Zhaoling and Longmen Binyang Central Cave.

Led by the Center for the Art of East Asia at UChicago and Xi’an Jiaotong University, with major museum partners in China, Japan, and the US.

https://caea.lib.uchicago.edu/dcadp/en/

#DigitalHeritage #EastAsianArt #3DReconstruction #MuseumTech #DigitalHumanities

GaVS: 3D-Grounded Video Stabilization via Temporally-Consistent Local Reconstruction
GaVS reformulates video stabilization as a 3D task: using temporally-consistent local reconstructions and Gaussian Splatting, it stabilizes full-frame video with high geometric consistency and robustness to motion and scene dynamics.
#SIGGRAPH2025 #VideoStabilization #3DReconstruction #GaussianSplatting
https://youtu.be/KNNdiv-Ea9g
https://sinoyou.github.io/gavs/
GaVS: 3D-Grounded Video Stabilization via Local Reconstruction and Rendering

YouTube
Top 7 Advances in 3D Reconstruction - Rackenzik

AI-driven 3D reconstruction enhances AR, VR, and robotics, enabling high-precision mapping and immersive digital experiences.

Rackenzik
Researchers Create 3D Model of Ancient Stone Sculpture From a Single 134-Year-Old Photo

They employed novel AI tech.

PetaPixel