Apple proposes HeadsUp for large-scale 3D Gaussian head reconstruction using 10k+ subject dataset
rsasaki0109 · x · 2026-08-16
Apple proposes HeadsUp, a scalable feed-forward method for reconstructing high-quality 3D Gaussian heads from multi-camera setups.
It uses an encoder-decoder architecture to compress views into a latent representation, decoded into UV-parameterized 3D Gaussians anchored to a neutral template. This decouples Gaussian count from input resolution, enabling training with many high-res views.
Trained on an internal dataset of over 10,000 subjects—an order of magnitude larger than existing datasets—it achieves state-of-the-art reconstruction performance.
More from Multimodal
- Fan-made video with Minimax H3: Clawhauser asks Sonic about Amy in Zootopia style — TigerClaw305 · 2026-08-16
- Seedance 2.5 launches on TapNow with native 1080p video generation — SimplyAnnisa · 2026-08-16
- AtomicDance Generates Dances Using Atomic Movements — jiqizhixin · 2026-08-16
- Ultimate SD Upscale with MiniMax H3: 2K Image in 25 Mins on 16GB VRAM — alisitskii · 2026-08-16
- MiniMax H3 integration with ComfyUI USDU workflow benchmarked — alisitskii · 2026-08-16
- Grok 4.6 Helps Build Minecraft Clone for Kids, Showcasing Game Generation — minchoi · 2026-08-16