FlowIBR: Leveraging Pre-Training for Efficient Neural Image-Based Rendering of Dynamic Scenes
This addresses the efficiency bottleneck for dynamic scene rendering in computer vision, enabling faster processing on consumer hardware, though it is incremental as it builds on existing neural image-based rendering methods.
The paper tackles the problem of slow optimization times in monocular novel view synthesis for dynamic scenes by introducing FlowIBR, which integrates pre-training on static scenes with per-scene flow optimization, reducing per-scene optimization time by an order of magnitude while maintaining comparable rendering quality.
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.