Live Fish Species Classification in Underwater Images by Using Convolutional Neural Networks Based on Incremental Learning with Knowledge Distillation Loss
Nowadays, underwater video systems are largely used by marine ecologists to study the biodiversity in underwater environments. These systems are non-destructive, do not perturb the environment and generate a large amount of visual data usable at any time. However, automatic video analysis requires e...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-08-01
|
Series: | Machine Learning and Knowledge Extraction |
Subjects: | |
Online Access: | https://www.mdpi.com/2504-4990/4/3/36 |
_version_ | 1797485476866686976 |
---|---|
author | Abdelouahid Ben Tamou Abdesslam Benzinou Kamal Nasreddine |
author_facet | Abdelouahid Ben Tamou Abdesslam Benzinou Kamal Nasreddine |
author_sort | Abdelouahid Ben Tamou |
collection | DOAJ |
description | Nowadays, underwater video systems are largely used by marine ecologists to study the biodiversity in underwater environments. These systems are non-destructive, do not perturb the environment and generate a large amount of visual data usable at any time. However, automatic video analysis requires efficient techniques of image processing due to the poor quality of underwater images and the challenging underwater environment. In this paper, we address live reef fish species classification in an unconstrained underwater environment. We propose using a deep Convolutional Neural Network (CNN) and training this network by using a new strategy based on incremental learning. This training strategy consists of training the CNN progressively by focusing at first on learning the difficult species well and then gradually learning the new species incrementally using knowledge distillation loss while keeping the high performances of the old species already learned. The proposed approach yields an accuracy of 81.83% on the LifeClef 2015 Fish benchmark dataset. |
first_indexed | 2024-03-09T23:20:18Z |
format | Article |
id | doaj.art-37a2a773d3dd41b5b926c1bc13de25ad |
institution | Directory Open Access Journal |
issn | 2504-4990 |
language | English |
last_indexed | 2024-03-09T23:20:18Z |
publishDate | 2022-08-01 |
publisher | MDPI AG |
record_format | Article |
series | Machine Learning and Knowledge Extraction |
spelling | doaj.art-37a2a773d3dd41b5b926c1bc13de25ad2023-11-23T17:28:14ZengMDPI AGMachine Learning and Knowledge Extraction2504-49902022-08-014375376710.3390/make4030036Live Fish Species Classification in Underwater Images by Using Convolutional Neural Networks Based on Incremental Learning with Knowledge Distillation LossAbdelouahid Ben Tamou0Abdesslam Benzinou1Kamal Nasreddine2ENIB, UMR CNRS 6285 LabSTICC, 29238 Brest, FranceENIB, UMR CNRS 6285 LabSTICC, 29238 Brest, FranceENIB, UMR CNRS 6285 LabSTICC, 29238 Brest, FranceNowadays, underwater video systems are largely used by marine ecologists to study the biodiversity in underwater environments. These systems are non-destructive, do not perturb the environment and generate a large amount of visual data usable at any time. However, automatic video analysis requires efficient techniques of image processing due to the poor quality of underwater images and the challenging underwater environment. In this paper, we address live reef fish species classification in an unconstrained underwater environment. We propose using a deep Convolutional Neural Network (CNN) and training this network by using a new strategy based on incremental learning. This training strategy consists of training the CNN progressively by focusing at first on learning the difficult species well and then gradually learning the new species incrementally using knowledge distillation loss while keeping the high performances of the old species already learned. The proposed approach yields an accuracy of 81.83% on the LifeClef 2015 Fish benchmark dataset.https://www.mdpi.com/2504-4990/4/3/36underwater imagefish recognitiondeep learningconvolutional neural networkincremental learningknowledge distillation loss |
spellingShingle | Abdelouahid Ben Tamou Abdesslam Benzinou Kamal Nasreddine Live Fish Species Classification in Underwater Images by Using Convolutional Neural Networks Based on Incremental Learning with Knowledge Distillation Loss Machine Learning and Knowledge Extraction underwater image fish recognition deep learning convolutional neural network incremental learning knowledge distillation loss |
title | Live Fish Species Classification in Underwater Images by Using Convolutional Neural Networks Based on Incremental Learning with Knowledge Distillation Loss |
title_full | Live Fish Species Classification in Underwater Images by Using Convolutional Neural Networks Based on Incremental Learning with Knowledge Distillation Loss |
title_fullStr | Live Fish Species Classification in Underwater Images by Using Convolutional Neural Networks Based on Incremental Learning with Knowledge Distillation Loss |
title_full_unstemmed | Live Fish Species Classification in Underwater Images by Using Convolutional Neural Networks Based on Incremental Learning with Knowledge Distillation Loss |
title_short | Live Fish Species Classification in Underwater Images by Using Convolutional Neural Networks Based on Incremental Learning with Knowledge Distillation Loss |
title_sort | live fish species classification in underwater images by using convolutional neural networks based on incremental learning with knowledge distillation loss |
topic | underwater image fish recognition deep learning convolutional neural network incremental learning knowledge distillation loss |
url | https://www.mdpi.com/2504-4990/4/3/36 |
work_keys_str_mv | AT abdelouahidbentamou livefishspeciesclassificationinunderwaterimagesbyusingconvolutionalneuralnetworksbasedonincrementallearningwithknowledgedistillationloss AT abdesslambenzinou livefishspeciesclassificationinunderwaterimagesbyusingconvolutionalneuralnetworksbasedonincrementallearningwithknowledgedistillationloss AT kamalnasreddine livefishspeciesclassificationinunderwaterimagesbyusingconvolutionalneuralnetworksbasedonincrementallearningwithknowledgedistillationloss |