Two-step Transfer Learning for Semantic Plant Segmentation
Two-step Transfer Learning for Semantic Plant Segmentation
复制标题
DOI：
10.5220/0006576303320339
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Shunsuke Sakurai;Hideaki Uchiyama;Atsushi Shimada;Daisaku Arita;R. Taniguchi
中科院分区：
文献类型：
--
作者：
Shunsuke Sakurai;Hideaki Uchiyama;Atsushi Shimada;Daisaku Arita;R. Taniguchi
我们讨论了全卷积网络（FCN）对植物分割任务的适用性，该网络在语义分割任务中提供了有前途的性能。挑战在于用小数据集训练网络，因为与 ImageNet 和 PASCAL VOC 数据集等对象图像数据集相比，植物图像数据集中的样本不多。所提出的方法受到迁移学习的启发，但涉及两步适应。第一步，我们将迁移学习从包含许多具有大量标记数据的对象的源域应用到植物域中的主要类别。然后，在第二步中，使用植物域内的一些样本从主要类别到次要类别进行类别适应。在叶分割挑战（LSC）数据集上，实验结果证实了该方法的有效性，例如，A2数据集的F测量标准为0.953，比直接适应高0.355，比非直接适应高0.527。
We discuss the applicability of a fully convolutional network (FCN), which provides promising performance in semantic segmentation tasks, to plant segmentation tasks. The challenge lies in training the network with a small dataset because there are not many samples in plant image datasets, as compared to object image datasets such as ImageNet and PASCAL VOC datasets. The proposed method is inspired by transfer learning, but involves a two-step adaptation. In the first step, we apply transfer learning from a source domain that contains many objects with a large amount of labeled data to a major category in the plant domain. Then, in the second step, category adaptation is performed from the major category to a minor category with a few samples within the plant domain. With leaf segmentation challenge (LSC) dataset, the experimental results confirm the effectiveness of the proposed method such that F-measure criterion was, for instance, 0.953 for the A2 dataset, which was 0.355 higher than that of direct adaptation, and 0.527 higher than that of non-