An open-set framework for underwater image classification using autoencoders
Article highlights Fish classification could be considered as an open-set problem, since the number of unknown underwater species is extremely high. Employing autoencoders, we can make the model to memorize the familiar species and throw away unfamiliar ones. The number of species highly affects the...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
Springer
2022-07-01
|
Series: | SN Applied Sciences |
Subjects: | |
Online Access: | https://doi.org/10.1007/s42452-022-05105-w |
Summary: | Article highlights Fish classification could be considered as an open-set problem, since the number of unknown underwater species is extremely high. Employing autoencoders, we can make the model to memorize the familiar species and throw away unfamiliar ones. The number of species highly affects the performance of the classification, while it roughly does not affect the false alarm rate of non underwater objects. |
---|---|
ISSN: | 2523-3963 2523-3971 |