FewShotNeRF: Meta-Learning-based Novel View Synthesis for Rapid Scene-Specific Adaptation
This work addresses the challenge of rapid scene-specific adaptation for novel view synthesis, which is incremental by building on existing NeRF and meta-learning techniques.
The paper tackles the problem of generating novel views of real-world objects with limited multi-view images by proposing FewShotNeRF, a meta-learning-based method that accelerates Neural Radiance Field adaptation to specific scenes, achieving high-quality novel view synthesis as demonstrated on the Common Objects in 3D dataset.
In this paper, we address the challenge of generating novel views of real-world objects with limited multi-view images through our proposed approach, FewShotNeRF. Our method utilizes meta-learning to acquire optimal initialization, facilitating rapid adaptation of a Neural Radiance Field (NeRF) to specific scenes. The focus of our meta-learning process is on capturing shared geometry and textures within a category, embedded in the weight initialization. This approach expedites the learning process of NeRFs and leverages recent advancements in positional encodings to reduce the time required for fitting a NeRF to a scene, thereby accelerating the inner loop optimization of meta-learning. Notably, our method enables meta-learning on a large number of 3D scenes to establish a robust 3D prior for various categories. Through extensive evaluations on the Common Objects in 3D open source dataset, we empirically demonstrate the efficacy and potential of meta-learning in generating high-quality novel views of objects.