| license: mit | |
| # FoundObj: Self-supervised Foundation Models as Rewards for Label-free 3D Object Segmentation | |
| FoundObj is an unsupervised 3D object segmentation framework with the aid of unsupervised foundation models. | |
| **Paper:** [arXiv:2605.27178](https://arxiv.org/abs/2605.27178) | |
| ## Contents | |
| This repository contains: | |
| - Pre-trained model checkpoint | |
| - Processed Center Field training data | |