Authors:
Shunsuke Sakurai
1
;
Hideaki Uchiyama
1
;
Atsushi Shimada
1
;
Daisaku Arita
2
and
Rin-ichiro Taniguchi
1
Affiliations:
1
Kyushu University, Japan
;
2
University of Nagasaki, Japan
Keyword(s):
Semantic Segmentation, Transfer Learning, Deep Learning, CNN, Plant Segmentation.
Related
Ontology
Subjects/Areas/Topics:
Knowledge Acquisition and Representation
;
Pattern Recognition
;
Theory and Methods
Abstract:
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 0.35
5 higher than that of direct adaptation, and 0.527 higher than that of non-adaptation.
(More)