ABLE-NeRF: Attention-Based Rendering with Learnable Embeddings for Neural Radiance Field
This work addresses rendering artifacts in 3D scene reconstruction for computer vision and graphics applications, representing an incremental improvement over existing NeRF methods.
The paper tackles the problem of rendering glossy and translucent surfaces in Neural Radiance Fields (NeRF), which often appear murky due to limitations in volumetric rendering, by introducing an attention-based framework with learnable embeddings. The result is a method called ABLE-NeRF that significantly reduces blurry surfaces and achieves state-of-the-art performance on the Blender dataset, surpassing Ref-NeRF in PSNR, SSIM, and LPIPS metrics.
Neural Radiance Field (NeRF) is a popular method in representing 3D scenes by optimising a continuous volumetric scene function. Its large success which lies in applying volumetric rendering (VR) is also its Achilles' heel in producing view-dependent effects. As a consequence, glossy and transparent surfaces often appear murky. A remedy to reduce these artefacts is to constrain this VR equation by excluding volumes with back-facing normal. While this approach has some success in rendering glossy surfaces, translucent objects are still poorly represented. In this paper, we present an alternative to the physics-based VR approach by introducing a self-attention-based framework on volumes along a ray. In addition, inspired by modern game engines which utilise Light Probes to store local lighting passing through the scene, we incorporate Learnable Embeddings to capture view dependent effects within the scene. Our method, which we call ABLE-NeRF, significantly reduces `blurry' glossy surfaces in rendering and produces realistic translucent surfaces which lack in prior art. In the Blender dataset, ABLE-NeRF achieves SOTA results and surpasses Ref-NeRF in all 3 image quality metrics PSNR, SSIM, LPIPS.