WildRayZer: Self-supervised Large View Synthesis in Dynamic Environments

TL;DR

WildRayZer is a self-supervised framework for novel view synthesis in dynamic environments, effectively handling moving objects and cameras by analyzing residuals to focus on background reconstruction, and is validated on a new large-scale real-world dataset.

Contribution

It introduces WildRayZer, a novel self-supervised approach for dynamic scene view synthesis that leverages residual analysis and pseudo motion masks, along with a new dataset for evaluation.

Findings

Outperforms existing methods in transient-region removal

Achieves higher quality full-frame NVS with a single feed-forward pass

Demonstrates robustness in dynamic, real-world scenarios

Abstract

We present WildRayZer, a self-supervised framework for novel view synthesis (NVS) in dynamic environments where both the camera and objects move. Dynamic content breaks the multi-view consistency that static NVS models rely on, leading to ghosting, hallucinated geometry, and unstable pose estimation. WildRayZer addresses this by performing an analysis-by-synthesis test: a camera-only static renderer explains rigid structure, and its residuals reveal transient regions. From these residuals, we construct pseudo motion masks, distill a motion estimator, and use it to mask input tokens and gate loss gradients so supervision focuses on cross-view background completion. To enable large-scale training and evaluation, we curate Dynamic RealEstate10K (D-RE10K), a real-world dataset of 15K casually captured dynamic sequences, and D-RE10K-iPhone, a paired transient and clean benchmark for sparse-view transient-aware NVS. Experiments show that WildRayZer consistently outperforms optimization-based and feed-forward baselines in both transient-region removal and full-frame NVS quality with a single feed-forward pass.