A Unified Point-Based Framework for 3D Segmentation
This work addresses segmentation problems in 3D computer vision for applications like robotics and AR/VR, representing an incremental improvement through feature integration and camera pose synthesis.
The paper tackles 3D point cloud segmentation challenges in structureless and textureless regions by proposing a unified point-based framework that optimizes pixel-level features, geometrical structures, and global context priors, outperforming state-of-the-art approaches on the ScanNet benchmark.
3D point cloud segmentation remains challenging for structureless and textureless regions. We present a new unified point-based framework for 3D point cloud segmentation that effectively optimizes pixel-level features, geometrical structures and global context priors of an entire scene. By back-projecting 2D image features into 3D coordinates, our network learns 2D textural appearance and 3D structural features in a unified framework. In addition, we investigate a global context prior to obtain a better prediction. We evaluate our framework on ScanNet online benchmark and show that our method outperforms several state-of-the-art approaches. We explore synthesizing camera poses in 3D reconstructed scenes for achieving higher performance. In-depth analysis on feature combinations and synthetic camera pose verify that features from different modalities benefit each other and dense camera pose sampling further improves the segmentation results.