FlowIBR: Leveraging Pre-Training for Efficient Neural Image-Based Rendering of Dynamic Scenes

TL;DR

FlowIBR is a new method that uses pre-training and scene flow to efficiently generate novel views of dynamic scenes, significantly reducing optimization time while maintaining high quality.

Contribution

It introduces FlowIBR, combining pre-trained image-based rendering with scene flow for fast dynamic scene rendering on a single GPU.

Findings

Reduces per-scene optimization time by an order of magnitude.

Achieves comparable rendering quality to existing methods.

Operates efficiently on a consumer-grade GPU.

Abstract

We introduce FlowIBR, a novel approach for efficient monocular novel view synthesis of dynamic scenes. Existing techniques already show impressive rendering quality but tend to focus on optimization within a single scene without leveraging prior knowledge, resulting in long optimization times per scene. FlowIBR circumvents this limitation by integrating a neural image-based rendering method, pre-trained on a large corpus of widely available static scenes, with a per-scene optimized scene flow field. Utilizing this flow field, we bend the camera rays to counteract the scene dynamics, thereby presenting the dynamic scene as if it were static to the rendering network. The proposed method reduces per-scene optimization time by an order of magnitude, achieving comparable rendering quality to existing methods -- all on a single consumer-grade GPU.