Improving 2D Feature Representations by 3D-Aware Fine-Tuning
This addresses the problem of limited 3D awareness in 2D foundation models for computer vision researchers, representing an incremental improvement through a novel fine-tuning method.
The paper tackles the limitation of 2D visual foundation models lacking 3D understanding by fine-tuning them on 3D-aware data, resulting in improved features that boost performance in semantic segmentation and depth estimation across multiple datasets.
Current visual foundation models are trained purely on unstructured 2D data, limiting their understanding of 3D structure of objects and scenes. In this work, we show that fine-tuning on 3D-aware data improves the quality of emerging semantic features. We design a method to lift semantic 2D features into an efficient 3D Gaussian representation, which allows us to re-render them for arbitrary views. Using the rendered 3D-aware features, we design a fine-tuning strategy to transfer such 3D awareness into a 2D foundation model. We demonstrate that models fine-tuned in that way produce features that readily improve downstream task performance in semantic segmentation and depth estimation through simple linear probing. Notably, though fined-tuned on a single indoor dataset, the improvement is transferable to a variety of indoor datasets and out-of-domain datasets. We hope our study encourages the community to consider injecting 3D awareness when training 2D foundation models. Project page: https://ywyue.github.io/FiT3D.