Journal
ICES JOURNAL OF MARINE SCIENCE
Volume 77, Issue 4, Pages 1427-1439Publisher
OXFORD UNIV PRESS
DOI: 10.1093/icesjms/fsz171
Keywords
deep learning; microscopic image; object detection; phytoplankton dataset
Categories
Funding
- National Natural Science Foundation of China [U1706218, 61971388]
- Key Research and Development Program of Shandong Province [GG201703140154]
- Applied Basic Research Programs of Qingdao [18-2-2-38-jch]
Ask authors/readers for more resources
Phytoplankton plays an important role in marine ecological environment and aquaculture. However, the recognition and detection of phytoplankton rely on manual operations. As the foundation of achieving intelligence and releasing human labour, a phytoplankton microscopic image dataset PMID2019 for phytoplankton automated detection is presented. The PMID2019 dataset contains 10819 phytoplankton microscopic images of 24 different categories. We leverage microscopes to collect images of phytoplankton in the laboratory environment. Each object in the images is manually labelled with a bounding box and category of ground-truth. In addition, living cells move quickly making it difficult to capture images of them. In order to generalize the dataset for in situ applications, we further utilize Cycle-GAN to achieve the domain migration between dead and living cell samples. We built a synthetic dataset to generate the corresponding living cell samples from the original dead ones. The PMID2019 dataset will not only benefit the development of phytoplankton microscopic vision technology in the future, but also can be widely used to assess the performance of the state-of-the-art object detection algorithms for phytoplankton recognition. Finally, we illustrate the performances of some state-of-the-art object detection algorithms, which may provide new ideas for monitoring marine ecosystems.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available