Efficient and accurate 3D reconstruction is crucial for cultural heritage applications. This study addresses the challenge of visualizing objects in complex scenes at high levels of detail (LOD) using Neural Radiance Fields (NeRFs), improving visual fidelity for selected objects while maintaining computational efficiency. The proposed ROI-NeRFs framework divides the scene into a Scene NeRF, capturing the overall scene at moderate detail, and multiple Region Of Interest (ROI) NeRFs, focusing on user-defined objects. An object-focused camera selection module automatically groups relevant cameras for each NeRF during the decomposition phase. In contrast, a Ray-level Compositional Rendering technique in the composition phase integrates Scene and ROI NeRFs for simultaneous multi-object rendering. Experiments on two real-world datasets, including a complex eighteenth-century cultural heritage room, demonstrate superior performance over baseline methods, enhancing LOD in object regions, minimizing artifacts, and maintaining efficient inference.

3D reconstruction of cultural heritage sites is essential for creating digital twins used in archiving, conservation [Kong and Hucks, 2023], archaeology [Haibt, 2024], and interactive museum exhibitions [Liu and Chang, 2024].

Neural Radiance Field (NeRF) [Mildenhall et al., 2021] has emerged as a powerful alternative for novel view synthesis (NVS) with unprecedented image quality. NeRF learns scenes as continuous fields of volume density and view-dependent color using MLPs, achieving photorealistic rendering through differentiable volume rendering. However, most NeRF-based methods train a single model on all images of a scene, yielding renderings with a uniform level of detail (LOD). In large-scale scenes, this fixed resolution might lead to poor-quality when viewing close-up specific objects or regions of interest (ROI), while increasing detail globally would incur prohibitive computational costs.