Skip to main content
GameDev.net gamedev.net

Cvpr 2026 News

The latest Cvpr 2026 coverage curated for game developers.

Dense Metric Depth Completion from Sparse Direct Time-of-Flight Sensors

A CVPR 2026 paper tackles a practical depth problem for VR, robotics, and 3D perception: turning sparse direct time-of-flight readings into dense metric depth. The system uses a dual-branch Vision …

Research arXiv cs.GR · 1 month, 3 weeks ago
What Is It Like to Be a Noise? An Entropy-based Gaussian Noise Regularization for Diffusion Models

Disney Research Studios is tackling a subtle failure mode in diffusion-model latent optimization: the sample can stop looking like real Gaussian noise even when the objective improves. Their fix is …

Research Disney Research Studios · 3 months, 4 weeks ago
RenderFlow: Single-Step Neural Rendering via Flow Matching

Disney Research’s RenderFlow replaces iterative diffusion with a deterministic single-step neural renderer, aiming for near real-time output without the usual stochastic flicker. For graphics teams, the interesting part is the …

Research Disney Research Studios · 3 months, 4 weeks ago
RelightAnyone: A Generalized Relightable 3D Gaussian Head Model

Disney Research’s new head-relighting model removes a big capture bottleneck: you no longer need OLAT footage for every subject. It can learn from ordinary single- or multi-view images, then generalize …

Research Disney Research Studios · 3 months, 4 weeks ago
Multi-view Consistent 3D Gaussian Head Avatars 'without' Multi-view Generation

A paper from arXiv claims it can train high-fidelity 3D Gaussian head avatars from random 2D images alone, skipping multi-view capture, 3D supervision, and intermediate view synthesis. For teams building …

Research arXiv cs.GR · 4 months ago