Title

Multi-species seagrass detection using semi-supervised learning

Document Type

Conference Proceeding

Publication Title

2021 36th International Conference on Image and Vision Computing New Zealand (IVCNZ)

Volume

2021-December

Publisher

IEEE

School

School of Science / Centre for Marine Ecosystems Research / Graduate Research

RAS ID

40652

Comments

Noman, M. K., Islam, S. M. S., Abu-Khalaf, J., & Lavery, P. (2021, December). Multi-species seagrass detection using semi-supervised learning [Paper presentation]. 2021 36th International Conference on Image and Vision Computing New Zealand (IVCNZ), Tauranga, New Zealand.

https://doi.org/10.1109/IVCNZ54163.2021.9653222

Abstract

This paper introduces a semi-supervised two-stage framework that leverages a large collection of unlabelled seagrass data with the guidance of a small amount of labelled seagrass data. First, a state-of-the-art Convolutional Neural Network (CNN) named EfficientNet is trained on a small number of labelled seagrass image patches. Next, the trained classifier generates pseudo labels for the unlabelled seagrass image patches. Finally, this classifier is retrained on the combination of labelled and pseudo labelled seagrass image patches. This approach achieves 98.0% multi-species detection accuracy on a newly developed seagrass dataset. Our approach on the publicly available 'DeepSeagrass' dataset achieves an overall accuracy of 91.3% and 93.0%, which outperforms the state-of-the-art multi-species seagrass detection accuracy of 88.2% and 92.4% for four and five classes, respectively.

DOI

10.1109/IVCNZ54163.2021.9653222

Access Rights

subscription content

Research Themes

Natural and Built Environments

Priority Areas

Environmental management, governance and policy

Share

 
COinS