Learning to Group: A Bottom-Up Framework for 3D Part Discovery in Unseen Categories
This addresses the challenge of generalizing 3D part segmentation to new object categories, which is incremental as it builds on existing shape segmentation methods.
The paper tackles the problem of discovering 3D parts for objects in unseen categories by proposing a learning-based agglomerative clustering framework, achieving state-of-the-art performance on the PartNet dataset by transferring knowledge from 3 training categories to 21 unseen testing categories without annotated samples.
We address the problem of discovering 3D parts for objects in unseen categories. Being able to learn the geometry prior of parts and transfer this prior to unseen categories pose fundamental challenges on data-driven shape segmentation approaches. Formulated as a contextual bandit problem, we propose a learning-based agglomerative clustering framework which learns a grouping policy to progressively group small part proposals into bigger ones in a bottom-up fashion. At the core of our approach is to restrict the local context for extracting part-level features, which encourages the generalizability to unseen categories. On the large-scale fine-grained 3D part dataset, PartNet, we demonstrate that our method can transfer knowledge of parts learned from 3 training categories to 21 unseen testing categories without seeing any annotated samples. Quantitative comparisons against four shape segmentation baselines shows that our approach achieve the state-of-the-art performance.