Unsupervised Knowledge Transfer for Object Detection in Marine Environmental Monitoring and Exploration
Open Access
- 5 August 2020
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 8, 143558-143568
Abstract
The volume of digital image data collected in the field of marine environmental monitoring and exploration has been growing in rapidly increasing rates in recent years. Computational support is essential for the timely evaluation of the high volume of marine imaging data, but often modern techniques such as deep learning cannot be applied due to the lack of training data. In this article, we present Unsupervised Knowledge Transfer (UnKnoT), a new method to use the limited amount of training data more efficiently. In order to avoid time-consuming annotation, it employs a technique we call “scale transfer” and enhanced data augmentation to reuse existing training data for object detection of the same object classes in new image datasets. We introduce four fully annotated marine image datasets acquired in the same geographical area but with different gear and distance to the sea floor. We evaluate the new method on the four datasets and show that it can greatly improve the object detection performance in the relevant cases compared to object detection without knowledge transfer. We conclude with a recommendation for an image acquisition and annotation scheme that ensures a good applicability of modern machine learning methods in the field of marine environmental monitoring and exploration.Keywords
Funding Information
- BMBF Project COSEMIO (FKZ 03F0812C)
- BMBF-funded de.NBI Cloud within the German Network for Bioinformatics Infrastructure (031A537B, 031A533A, 031A538A, 031A533B, 031A535A, 031A537C, 031A534A, 031A532B)
- German Research Foundation
- Open Access Publication Fund of Bielefeld University
This publication has 30 references indexed in Scilit:
- Comparison of image annotation data generated by multiple investigators for benthic ecologyMarine Ecology Progress Series, 2016
- RecoMIA—Recommendations for Marine Image Annotation: Lessons Learned and Future DirectionsFrontiers in Marine Science, 2016
- DELPHI—fast and adaptive computational laser point detection and visual footprint quantification for arbitrary underwater image collectionsFrontiers in Marine Science, 2015
- Semi-Automated Image Analysis for the Assessment of Megafaunal Densities at the Arctic Deep-Sea Observatory HAUSGARTENPLOS ONE, 2012
- Automated annotation of coral reef survey images2012
- A Survey on Transfer LearningIEEE Transactions on Knowledge and Data Engineering, 2009
- The Pascal Visual Object Classes (VOC) ChallengeInternational Journal of Computer Vision, 2009
- A systematic analysis of performance measures for classification tasksInformation Processing & Management, 2009
- VIPS: an image processing system for large images1996
- DISCOL: A Long-Term, Large-Scale, Disturbance-Recolonization Experiment in the Abyssal Eastern Tropical South Pacific Ocean1990