Skip to content
Tagged articles

novel view synthesis

6 articles · Page 1 of 1
Machine Heart
Machine Heart
Oct 19, 2026 · Artificial Intelligence

DynStream: First Pose-Free Streaming 4D Gaussian Reconstruction for Large-Scale Dynamic Scenes

Peking University, VGI Labs, and UC Merced introduce DynStream, a feed-forward framework that reconstructs dynamic 3D scenes from pose-free video streams by chunking input, predicting relative poses and scale factors for global alignment, and separating static and dynamic Gaussians for continuous rendering.

4D reconstructionComputer VisionGaussian Splatting
0 likes · 9 min read
DynStream: First Pose-Free Streaming 4D Gaussian Reconstruction for Large-Scale Dynamic Scenes
Machine Heart
Machine Heart
Oct 15, 2026 · Artificial Intelligence

Honeycomb: Fixed-Size Scene Memory for Consistent Long Video Generation

Honeycomb introduces HexMemory, a constant-size scene memory using six fixed feature planes that enables video world models to maintain long-term spatial consistency without growing memory footprint, outperforming prior methods in loop-closure revisit tasks and novel view synthesis.

HexMemoryconstant-size memorylong video generation
0 likes · 16 min read
Honeycomb: Fixed-Size Scene Memory for Consistent Long Video Generation
Machine Heart
Machine Heart
Jul 27, 2026 · Artificial Intelligence

UniWorld-View Tops WorldScore, Letting AI Infer Unseen Spaces and Create New View Videos

UniWorld-View advances world‑model research by introducing occlusion‑aware point‑cloud rendering and a dual‑stream conditioning architecture that lets AI infer unseen 3D geometry from a single image or video and generate high‑quality novel‑view videos, earning the top spot on the WorldScore leaderboard.

AI Video GenerationPoint Cloud RenderingUniWorld-View
0 likes · 11 min read
UniWorld-View Tops WorldScore, Letting AI Infer Unseen Spaces and Create New View Videos
JD Cloud Developers
JD Cloud Developers
Apr 22, 2025 · Artificial Intelligence

How AI Turns 2D Videos into Immersive 3D Spatial Content at Scale

Leveraging 3D vision and AIGC, JD Retail’s R&D team converts abundant 2D video assets into high‑quality stereoscopic 3D space videos through a pipeline that includes monocular depth estimation, novel view synthesis, multi‑branch inpainting, and MV‑HEVC encoding, validated by ICME 2025 and a new StereoV1K dataset.

3D videoAIGCComputer Vision
0 likes · 26 min read
How AI Turns 2D Videos into Immersive 3D Spatial Content at Scale
JD Tech Talk
JD Tech Talk
Apr 22, 2025 · Artificial Intelligence

End-to-End 3D Spatial Video Generation via Monocular Depth Estimation, Novel View Synthesis, and MV-HEVC Encoding

Leveraging AI-driven monocular depth estimation, novel view synthesis, and MV‑HEVC encoding, the JD Retail Content R&D team presents an end‑to‑end pipeline that converts 2D video assets into high‑quality immersive 3D spatial videos, introduces the large‑scale StereoV1K dataset, and demonstrates superior performance over existing methods.

3D video generationAIGCComputer Vision
0 likes · 22 min read
End-to-End 3D Spatial Video Generation via Monocular Depth Estimation, Novel View Synthesis, and MV-HEVC Encoding
JD Tech
JD Tech
Apr 21, 2025 · Artificial Intelligence

End-to-End 3D Spatial Video Generation via Monocular Depth Estimation, Novel View Synthesis, and MV‑HEVC Encoding

This article presents a comprehensive AI‑driven pipeline that converts 2D video into immersive 3D spatial video by leveraging monocular depth estimation, depth‑warping novel view synthesis, a multi‑branch inpainting module, a large‑scale StereoV1K dataset, and efficient MV‑HEVC compression, with results validated at ICME 2025 and deployed in JD Vision services.

3D videoAIAIGC
0 likes · 20 min read
End-to-End 3D Spatial Video Generation via Monocular Depth Estimation, Novel View Synthesis, and MV‑HEVC Encoding