PS4PRO: Pixel-to-pixel Supervision for Photorealistic Rendering and Optimization
This addresses the challenge of reconstructing complex and dynamic 3D scenes from insufficient 2D images, which is incremental as it builds on existing neural rendering methods with a data augmentation approach.
The paper tackles the problem of limited input views in neural rendering for 3D scene reconstruction by proposing PS4PRO, a lightweight video frame interpolation model that augments datasets, resulting in improved reconstruction performance on static and dynamic scenes.
Neural rendering methods have gained significant attention for their ability to reconstruct 3D scenes from 2D images. The core idea is to take multiple views as input and optimize the reconstructed scene by minimizing the uncertainty in geometry and appearance across the views. However, the reconstruction quality is limited by the number of input views. This limitation is further pronounced in complex and dynamic scenes, where certain angles of objects are never seen. In this paper, we propose to use video frame interpolation as the data augmentation method for neural rendering. Furthermore, we design a lightweight yet high-quality video frame interpolation model, PS4PRO (Pixel-to-pixel Supervision for Photorealistic Rendering and Optimization). PS4PRO is trained on diverse video datasets, implicitly modeling camera movement as well as real-world 3D geometry. Our model performs as an implicit world prior, enriching the photo supervision for 3D reconstruction. By leveraging the proposed method, we effectively augment existing datasets for neural rendering methods. Our experimental results indicate that our method improves the reconstruction performance on both static and dynamic scenes.