Multi-species Fish Identification using Hybrid DeepCNN with Refined Squeeze and Excitation Architecture
Fish play a prominent role in the food web and fish farming has value for both human consumption and tourist attractions. Due to the increasing importance of marine biodiversity, recognition of fish species has become a prominent task in monitoring the mislabelling of seafood and extinct species. Th...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Istanbul University Press
2022-10-01
|
| Series: | Aquatic Sciences and Engineering |
| Subjects: | |
| Online Access: | https://dergipark.org.tr/tr/download/article-file/2601117 |
| _version_ | 1797271986559254528 |
|---|---|
| author | Nivetha P Jansi Rani Sella Veluswami |
| author_facet | Nivetha P Jansi Rani Sella Veluswami |
| author_sort | Nivetha P |
| collection | DOAJ |
| description | Fish play a prominent role in the food web and fish farming has value for both human consumption and tourist attractions. Due to the increasing importance of marine biodiversity, recognition of fish species has become a prominent task in monitoring the mislabelling of seafood and extinct species. This problem can be solved using traditional manual annotation on the images. To reduce manpow-er, cost, and tremendous time, deep learning approaches are used which always require large data-sets. Therefore, fish species identification is a challenging task using disproportionately small data sets. In this research, we develop a new method by refining the squeeze and excitation network for the automatic fish species classification model to identify 23 different types of fish species. To achieve this, a hybrid framework using deep learning is proposed on a large-scale dataset and implemented transfer learning for a small-scale dataset. Deep learning methods can be used to identify fish in un-derwater images. In this study, we have proposed a new method of hybrid Deep Convolutional Neu-ralNetwork(CNN)alongwithaSupportVectorMachine(SVM)forclassification.Additionally,theSqueeze and Excitation (SE) block has been improved for improved feature extraction. The proposed method achieved an accuracy of 97.90%. Then post-training with the small-scale dataset (Croatian) achieved an accuracy of 94.99% with an 11% improvement compared to Bilinear CNN (B-CNN) (Qui et al., 2018) and can be used in any underwater applications to identify fish species and avoid misla-belling of seafood. |
| first_indexed | 2024-03-07T14:18:04Z |
| format | Article |
| id | doaj.art-029012e2cb2d48f8860873bf00a71349 |
| institution | Directory Open Access Journal |
| issn | 2602-473X |
| language | English |
| last_indexed | 2024-03-07T14:18:04Z |
| publishDate | 2022-10-01 |
| publisher | Istanbul University Press |
| record_format | Article |
| series | Aquatic Sciences and Engineering |
| spelling | doaj.art-029012e2cb2d48f8860873bf00a713492024-03-06T10:11:34ZengIstanbul University PressAquatic Sciences and Engineering2602-473X2022-10-0137422022810.26650/ASE2022211632024Multi-species Fish Identification using Hybrid DeepCNN with Refined Squeeze and Excitation ArchitectureNivetha P0Jansi Rani Sella Veluswami1Sri Sivasubramaniya Nadar College of EngineeringSri Sivasubramaniya Nadar College of EngineeringFish play a prominent role in the food web and fish farming has value for both human consumption and tourist attractions. Due to the increasing importance of marine biodiversity, recognition of fish species has become a prominent task in monitoring the mislabelling of seafood and extinct species. This problem can be solved using traditional manual annotation on the images. To reduce manpow-er, cost, and tremendous time, deep learning approaches are used which always require large data-sets. Therefore, fish species identification is a challenging task using disproportionately small data sets. In this research, we develop a new method by refining the squeeze and excitation network for the automatic fish species classification model to identify 23 different types of fish species. To achieve this, a hybrid framework using deep learning is proposed on a large-scale dataset and implemented transfer learning for a small-scale dataset. Deep learning methods can be used to identify fish in un-derwater images. In this study, we have proposed a new method of hybrid Deep Convolutional Neu-ralNetwork(CNN)alongwithaSupportVectorMachine(SVM)forclassification.Additionally,theSqueeze and Excitation (SE) block has been improved for improved feature extraction. The proposed method achieved an accuracy of 97.90%. Then post-training with the small-scale dataset (Croatian) achieved an accuracy of 94.99% with an 11% improvement compared to Bilinear CNN (B-CNN) (Qui et al., 2018) and can be used in any underwater applications to identify fish species and avoid misla-belling of seafood.https://dergipark.org.tr/tr/download/article-file/2601117deep learningconvolutional neural networksqueeze and excitationfish speciesfish4knowledge dataset |
| spellingShingle | Nivetha P Jansi Rani Sella Veluswami Multi-species Fish Identification using Hybrid DeepCNN with Refined Squeeze and Excitation Architecture Aquatic Sciences and Engineering deep learning convolutional neural network squeeze and excitation fish species fish4knowledge dataset |
| title | Multi-species Fish Identification using Hybrid DeepCNN with Refined Squeeze and Excitation Architecture |
| title_full | Multi-species Fish Identification using Hybrid DeepCNN with Refined Squeeze and Excitation Architecture |
| title_fullStr | Multi-species Fish Identification using Hybrid DeepCNN with Refined Squeeze and Excitation Architecture |
| title_full_unstemmed | Multi-species Fish Identification using Hybrid DeepCNN with Refined Squeeze and Excitation Architecture |
| title_short | Multi-species Fish Identification using Hybrid DeepCNN with Refined Squeeze and Excitation Architecture |
| title_sort | multi species fish identification using hybrid deepcnn with refined squeeze and excitation architecture |
| topic | deep learning convolutional neural network squeeze and excitation fish species fish4knowledge dataset |
| url | https://dergipark.org.tr/tr/download/article-file/2601117 |
| work_keys_str_mv | AT nivethap multispeciesfishidentificationusinghybriddeepcnnwithrefinedsqueezeandexcitationarchitecture AT jansiranisellaveluswami multispeciesfishidentificationusinghybriddeepcnnwithrefinedsqueezeandexcitationarchitecture |