PartDistillation: Learning Parts from Instance Segmentation
PartDistillation: Learning Parts from Instance Segmentation
复制标题
DOI：
10.1109/cvpr52729.2023.00691
复制
发表时间：
2023-06
期刊：
影响因子：
--
通讯作者：
Jang Hyun Cho
中科院分区：
文献类型：
--
作者：
Jang Hyun Cho
我们提出了一个可扩展的框架来学习对象实例标签的部分分割。最先进的实例分割模型包含数量惊人的零件信息。然而，这些信息中的大部分都隐藏在普通视图中。对于每个对象实例，零件信息都是嘈杂的、不一致的和不完整的。PartDistillation通过在大数据集上进行自监督自训练，将实例分割模型的部分信息转换为部分分割模型。所得到的分割模型是鲁棒的，准确的，以及推广。我们在各种部分分割数据集上评估模型。我们的模型优于监督部分分割在零杆泛化性能的大幅度提高。与监督对应物和其他基线相比，我们的模型在目标数据集上进行微调时表现出色，特别是在少数情况下。最后，我们的模型提供了一个更广泛的覆盖面罕见的部分时，超过10K的对象类进行评估。代码位于https://github.com/facebookresearch/PartDistillation。
We present a scalable framework to learn part segmentation from object instance labels. State-of-the-art instance segmentation models contain a surprising amount of part information. However, much of this information is hidden from plain view. For each object instance, the part information is noisy, inconsistent, and incomplete. PartDistillation transfers the part information of an instance segmentation model into a part segmentation model through self-supervised self-training on a large dataset. The resulting segmentation model is robust, accurate, and generalizes well. We evaluate the model on various part segmentation datasets. Our model outperforms supervised part segmentation in zero-shot generalization performance by a large margin. Our model outperforms when finetuned on target datasets compared to supervised counterpart and other baselines especially in few-shot regime. Finally, our model provides a wider coverage of rare parts when evaluated over 10K object classes. Code is at https://github.com/facebookresearch/PartDistillation.