Underwater Fish Species Identification Model and Real-Time Identification System
Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in ill...
Ausführliche Beschreibung
Gespeichert in:
| Autor*in: |
LI Shaobo [verfasserIn] YANG Ling [verfasserIn] YU Huihui [verfasserIn] CHEN Yingyi [verfasserIn] |
|---|
| Format: |
E-Artikel |
|---|---|
| Sprache: |
Englisch ; Chinesisch |
| Erschienen: |
2022 |
|---|
| Schlagwörter: |
|---|
| Übergeordnetes Werk: |
In: 智慧农业 - Editorial Office of Smart Agriculture, 2021, 4(2022), 1, Seite 130-139 |
|---|---|
| Übergeordnetes Werk: |
volume:4 ; year:2022 ; number:1 ; pages:130-139 |
| Links: |
|---|
| DOI / URN: |
10.12133/j.smartag.SA202202006 |
|---|
| Katalog-ID: |
DOAJ041492250 |
|---|
| LEADER | 01000caa a22002652 4500 | ||
|---|---|---|---|
| 001 | DOAJ041492250 | ||
| 003 | DE-627 | ||
| 005 | 20230503110409.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 230227s2022 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.12133/j.smartag.SA202202006 |2 doi | |
| 035 | |a (DE-627)DOAJ041492250 | ||
| 035 | |a (DE-599)DOAJ6ee1de60dd7f467384128fea694724ad | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
| 041 | |a eng |a chi | ||
| 050 | 0 | |a S1-972 | |
| 050 | 0 | |a T1-995 | |
| 100 | 0 | |a LI Shaobo |e verfasserin |4 aut | |
| 245 | 1 | 0 | |a Underwater Fish Species Identification Model and Real-Time Identification System |
| 264 | 1 | |c 2022 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 520 | |a Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. | ||
| 650 | 4 | |a fish identification model | |
| 650 | 4 | |a cnn | |
| 650 | 4 | |a model evaluation | |
| 650 | 4 | |a android | |
| 650 | 4 | |a ground-truth | |
| 650 | 4 | |a real-time identification system | |
| 653 | 0 | |a Agriculture (General) | |
| 653 | 0 | |a Technology (General) | |
| 700 | 0 | |a YANG Ling |e verfasserin |4 aut | |
| 700 | 0 | |a YU Huihui |e verfasserin |4 aut | |
| 700 | 0 | |a CHEN Yingyi |e verfasserin |4 aut | |
| 773 | 0 | 8 | |i In |t 智慧农业 |d Editorial Office of Smart Agriculture, 2021 |g 4(2022), 1, Seite 130-139 |w (DE-627)DOAJ078619246 |x 20968094 |7 nnns |
| 773 | 1 | 8 | |g volume:4 |g year:2022 |g number:1 |g pages:130-139 |
| 856 | 4 | 0 | |u https://doi.org/10.12133/j.smartag.SA202202006 |z kostenfrei |
| 856 | 4 | 0 | |u https://doaj.org/article/6ee1de60dd7f467384128fea694724ad |z kostenfrei |
| 856 | 4 | 0 | |u http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006 |z kostenfrei |
| 856 | 4 | 2 | |u https://doaj.org/toc/2096-8094 |y Journal toc |z kostenfrei |
| 912 | |a GBV_USEFLAG_A | ||
| 912 | |a SYSFLAG_A | ||
| 912 | |a GBV_DOAJ | ||
| 912 | |a SSG-OLC-PHA | ||
| 951 | |a AR | ||
| 952 | |d 4 |j 2022 |e 1 |h 130-139 | ||
| author_variant |
l s ls y l yl y h yh c y cy |
|---|---|
| matchkey_str |
article:20968094:2022----::newtrihpceietfctomdlnratm |
| hierarchy_sort_str |
2022 |
| callnumber-subject-code |
S |
| publishDate |
2022 |
| allfields |
10.12133/j.smartag.SA202202006 doi (DE-627)DOAJ041492250 (DE-599)DOAJ6ee1de60dd7f467384128fea694724ad DE-627 ger DE-627 rakwb eng chi S1-972 T1-995 LI Shaobo verfasserin aut Underwater Fish Species Identification Model and Real-Time Identification System 2022 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. fish identification model cnn model evaluation android ground-truth real-time identification system Agriculture (General) Technology (General) YANG Ling verfasserin aut YU Huihui verfasserin aut CHEN Yingyi verfasserin aut In 智慧农业 Editorial Office of Smart Agriculture, 2021 4(2022), 1, Seite 130-139 (DE-627)DOAJ078619246 20968094 nnns volume:4 year:2022 number:1 pages:130-139 https://doi.org/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/article/6ee1de60dd7f467384128fea694724ad kostenfrei http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/toc/2096-8094 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ SSG-OLC-PHA AR 4 2022 1 130-139 |
| spelling |
10.12133/j.smartag.SA202202006 doi (DE-627)DOAJ041492250 (DE-599)DOAJ6ee1de60dd7f467384128fea694724ad DE-627 ger DE-627 rakwb eng chi S1-972 T1-995 LI Shaobo verfasserin aut Underwater Fish Species Identification Model and Real-Time Identification System 2022 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. fish identification model cnn model evaluation android ground-truth real-time identification system Agriculture (General) Technology (General) YANG Ling verfasserin aut YU Huihui verfasserin aut CHEN Yingyi verfasserin aut In 智慧农业 Editorial Office of Smart Agriculture, 2021 4(2022), 1, Seite 130-139 (DE-627)DOAJ078619246 20968094 nnns volume:4 year:2022 number:1 pages:130-139 https://doi.org/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/article/6ee1de60dd7f467384128fea694724ad kostenfrei http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/toc/2096-8094 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ SSG-OLC-PHA AR 4 2022 1 130-139 |
| allfields_unstemmed |
10.12133/j.smartag.SA202202006 doi (DE-627)DOAJ041492250 (DE-599)DOAJ6ee1de60dd7f467384128fea694724ad DE-627 ger DE-627 rakwb eng chi S1-972 T1-995 LI Shaobo verfasserin aut Underwater Fish Species Identification Model and Real-Time Identification System 2022 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. fish identification model cnn model evaluation android ground-truth real-time identification system Agriculture (General) Technology (General) YANG Ling verfasserin aut YU Huihui verfasserin aut CHEN Yingyi verfasserin aut In 智慧农业 Editorial Office of Smart Agriculture, 2021 4(2022), 1, Seite 130-139 (DE-627)DOAJ078619246 20968094 nnns volume:4 year:2022 number:1 pages:130-139 https://doi.org/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/article/6ee1de60dd7f467384128fea694724ad kostenfrei http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/toc/2096-8094 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ SSG-OLC-PHA AR 4 2022 1 130-139 |
| allfieldsGer |
10.12133/j.smartag.SA202202006 doi (DE-627)DOAJ041492250 (DE-599)DOAJ6ee1de60dd7f467384128fea694724ad DE-627 ger DE-627 rakwb eng chi S1-972 T1-995 LI Shaobo verfasserin aut Underwater Fish Species Identification Model and Real-Time Identification System 2022 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. fish identification model cnn model evaluation android ground-truth real-time identification system Agriculture (General) Technology (General) YANG Ling verfasserin aut YU Huihui verfasserin aut CHEN Yingyi verfasserin aut In 智慧农业 Editorial Office of Smart Agriculture, 2021 4(2022), 1, Seite 130-139 (DE-627)DOAJ078619246 20968094 nnns volume:4 year:2022 number:1 pages:130-139 https://doi.org/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/article/6ee1de60dd7f467384128fea694724ad kostenfrei http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/toc/2096-8094 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ SSG-OLC-PHA AR 4 2022 1 130-139 |
| allfieldsSound |
10.12133/j.smartag.SA202202006 doi (DE-627)DOAJ041492250 (DE-599)DOAJ6ee1de60dd7f467384128fea694724ad DE-627 ger DE-627 rakwb eng chi S1-972 T1-995 LI Shaobo verfasserin aut Underwater Fish Species Identification Model and Real-Time Identification System 2022 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. fish identification model cnn model evaluation android ground-truth real-time identification system Agriculture (General) Technology (General) YANG Ling verfasserin aut YU Huihui verfasserin aut CHEN Yingyi verfasserin aut In 智慧农业 Editorial Office of Smart Agriculture, 2021 4(2022), 1, Seite 130-139 (DE-627)DOAJ078619246 20968094 nnns volume:4 year:2022 number:1 pages:130-139 https://doi.org/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/article/6ee1de60dd7f467384128fea694724ad kostenfrei http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006 kostenfrei https://doaj.org/toc/2096-8094 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ SSG-OLC-PHA AR 4 2022 1 130-139 |
| language |
English Chinese |
| source |
In 智慧农业 4(2022), 1, Seite 130-139 volume:4 year:2022 number:1 pages:130-139 |
| sourceStr |
In 智慧农业 4(2022), 1, Seite 130-139 volume:4 year:2022 number:1 pages:130-139 |
| format_phy_str_mv |
Article |
| institution |
findex.gbv.de |
| topic_facet |
fish identification model cnn model evaluation android ground-truth real-time identification system Agriculture (General) Technology (General) |
| isfreeaccess_bool |
true |
| container_title |
智慧农业 |
| authorswithroles_txt_mv |
LI Shaobo @@aut@@ YANG Ling @@aut@@ YU Huihui @@aut@@ CHEN Yingyi @@aut@@ |
| publishDateDaySort_date |
2022-01-01T00:00:00Z |
| hierarchy_top_id |
DOAJ078619246 |
| id |
DOAJ041492250 |
| language_de |
englisch chinesisch |
| fullrecord |
<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01000caa a22002652 4500</leader><controlfield tag="001">DOAJ041492250</controlfield><controlfield tag="003">DE-627</controlfield><controlfield tag="005">20230503110409.0</controlfield><controlfield tag="007">cr uuu---uuuuu</controlfield><controlfield tag="008">230227s2022 xx |||||o 00| ||eng c</controlfield><datafield tag="024" ind1="7" ind2=" "><subfield code="a">10.12133/j.smartag.SA202202006</subfield><subfield code="2">doi</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-627)DOAJ041492250</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-599)DOAJ6ee1de60dd7f467384128fea694724ad</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-627</subfield><subfield code="b">ger</subfield><subfield code="c">DE-627</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1=" " ind2=" "><subfield code="a">eng</subfield><subfield code="a">chi</subfield></datafield><datafield tag="050" ind1=" " ind2="0"><subfield code="a">S1-972</subfield></datafield><datafield tag="050" ind1=" " ind2="0"><subfield code="a">T1-995</subfield></datafield><datafield tag="100" ind1="0" ind2=" "><subfield code="a">LI Shaobo</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Underwater Fish Species Identification Model and Real-Time Identification System</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="c">2022</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">Text</subfield><subfield code="b">txt</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">Computermedien</subfield><subfield code="b">c</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">Online-Ressource</subfield><subfield code="b">cr</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s.</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">fish identification model</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">cnn</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">model evaluation</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">android</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">ground-truth</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">real-time identification system</subfield></datafield><datafield tag="653" ind1=" " ind2="0"><subfield code="a">Agriculture (General)</subfield></datafield><datafield tag="653" ind1=" " ind2="0"><subfield code="a">Technology (General)</subfield></datafield><datafield tag="700" ind1="0" ind2=" "><subfield code="a">YANG Ling</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="0" ind2=" "><subfield code="a">YU Huihui</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="0" ind2=" "><subfield code="a">CHEN Yingyi</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="773" ind1="0" ind2="8"><subfield code="i">In</subfield><subfield code="t">智慧农业</subfield><subfield code="d">Editorial Office of Smart Agriculture, 2021</subfield><subfield code="g">4(2022), 1, Seite 130-139</subfield><subfield code="w">(DE-627)DOAJ078619246</subfield><subfield code="x">20968094</subfield><subfield code="7">nnns</subfield></datafield><datafield tag="773" ind1="1" ind2="8"><subfield code="g">volume:4</subfield><subfield code="g">year:2022</subfield><subfield code="g">number:1</subfield><subfield code="g">pages:130-139</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doi.org/10.12133/j.smartag.SA202202006</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doaj.org/article/6ee1de60dd7f467384128fea694724ad</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="2"><subfield code="u">https://doaj.org/toc/2096-8094</subfield><subfield code="y">Journal toc</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_USEFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SYSFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_DOAJ</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SSG-OLC-PHA</subfield></datafield><datafield tag="951" ind1=" " ind2=" "><subfield code="a">AR</subfield></datafield><datafield tag="952" ind1=" " ind2=" "><subfield code="d">4</subfield><subfield code="j">2022</subfield><subfield code="e">1</subfield><subfield code="h">130-139</subfield></datafield></record></collection>
|
| callnumber-first |
S - Agriculture |
| author |
LI Shaobo |
| spellingShingle |
LI Shaobo misc S1-972 misc T1-995 misc fish identification model misc cnn misc model evaluation misc android misc ground-truth misc real-time identification system misc Agriculture (General) misc Technology (General) Underwater Fish Species Identification Model and Real-Time Identification System |
| authorStr |
LI Shaobo |
| ppnlink_with_tag_str_mv |
@@773@@(DE-627)DOAJ078619246 |
| format |
electronic Article |
| delete_txt_mv |
keep |
| author_role |
aut aut aut aut |
| collection |
DOAJ |
| remote_str |
true |
| callnumber-label |
S1-972 |
| illustrated |
Not Illustrated |
| issn |
20968094 |
| topic_title |
S1-972 T1-995 Underwater Fish Species Identification Model and Real-Time Identification System fish identification model cnn model evaluation android ground-truth real-time identification system |
| topic |
misc S1-972 misc T1-995 misc fish identification model misc cnn misc model evaluation misc android misc ground-truth misc real-time identification system misc Agriculture (General) misc Technology (General) |
| topic_unstemmed |
misc S1-972 misc T1-995 misc fish identification model misc cnn misc model evaluation misc android misc ground-truth misc real-time identification system misc Agriculture (General) misc Technology (General) |
| topic_browse |
misc S1-972 misc T1-995 misc fish identification model misc cnn misc model evaluation misc android misc ground-truth misc real-time identification system misc Agriculture (General) misc Technology (General) |
| format_facet |
Elektronische Aufsätze Aufsätze Elektronische Ressource |
| format_main_str_mv |
Text Zeitschrift/Artikel |
| carriertype_str_mv |
cr |
| hierarchy_parent_title |
智慧农业 |
| hierarchy_parent_id |
DOAJ078619246 |
| hierarchy_top_title |
智慧农业 |
| isfreeaccess_txt |
true |
| familylinks_str_mv |
(DE-627)DOAJ078619246 |
| title |
Underwater Fish Species Identification Model and Real-Time Identification System |
| ctrlnum |
(DE-627)DOAJ041492250 (DE-599)DOAJ6ee1de60dd7f467384128fea694724ad |
| title_full |
Underwater Fish Species Identification Model and Real-Time Identification System |
| author_sort |
LI Shaobo |
| journal |
智慧农业 |
| journalStr |
智慧农业 |
| callnumber-first-code |
S |
| lang_code |
eng chi |
| isOA_bool |
true |
| recordtype |
marc |
| publishDateSort |
2022 |
| contenttype_str_mv |
txt |
| container_start_page |
130 |
| author_browse |
LI Shaobo YANG Ling YU Huihui CHEN Yingyi |
| container_volume |
4 |
| class |
S1-972 T1-995 |
| format_se |
Elektronische Aufsätze |
| author-letter |
LI Shaobo |
| doi_str_mv |
10.12133/j.smartag.SA202202006 |
| author2-role |
verfasserin |
| title_sort |
underwater fish species identification model and real-time identification system |
| callnumber |
S1-972 |
| title_auth |
Underwater Fish Species Identification Model and Real-Time Identification System |
| abstract |
Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. |
| abstractGer |
Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. |
| abstract_unstemmed |
Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s. |
| collection_details |
GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ SSG-OLC-PHA |
| container_issue |
1 |
| title_short |
Underwater Fish Species Identification Model and Real-Time Identification System |
| url |
https://doi.org/10.12133/j.smartag.SA202202006 https://doaj.org/article/6ee1de60dd7f467384128fea694724ad http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006 https://doaj.org/toc/2096-8094 |
| remote_bool |
true |
| author2 |
YANG Ling YU Huihui CHEN Yingyi |
| author2Str |
YANG Ling YU Huihui CHEN Yingyi |
| ppnlink |
DOAJ078619246 |
| callnumber-subject |
S - General Agriculture |
| mediatype_str_mv |
c |
| isOA_txt |
true |
| hochschulschrift_bool |
false |
| doi_str |
10.12133/j.smartag.SA202202006 |
| callnumber-a |
S1-972 |
| up_date |
2024-07-03T20:32:07.300Z |
| _version_ |
1803591328518373376 |
| fullrecord_marcxml |
<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01000caa a22002652 4500</leader><controlfield tag="001">DOAJ041492250</controlfield><controlfield tag="003">DE-627</controlfield><controlfield tag="005">20230503110409.0</controlfield><controlfield tag="007">cr uuu---uuuuu</controlfield><controlfield tag="008">230227s2022 xx |||||o 00| ||eng c</controlfield><datafield tag="024" ind1="7" ind2=" "><subfield code="a">10.12133/j.smartag.SA202202006</subfield><subfield code="2">doi</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-627)DOAJ041492250</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-599)DOAJ6ee1de60dd7f467384128fea694724ad</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-627</subfield><subfield code="b">ger</subfield><subfield code="c">DE-627</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1=" " ind2=" "><subfield code="a">eng</subfield><subfield code="a">chi</subfield></datafield><datafield tag="050" ind1=" " ind2="0"><subfield code="a">S1-972</subfield></datafield><datafield tag="050" ind1=" " ind2="0"><subfield code="a">T1-995</subfield></datafield><datafield tag="100" ind1="0" ind2=" "><subfield code="a">LI Shaobo</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Underwater Fish Species Identification Model and Real-Time Identification System</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="c">2022</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">Text</subfield><subfield code="b">txt</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">Computermedien</subfield><subfield code="b">c</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">Online-Ressource</subfield><subfield code="b">cr</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">Convolutional neural network models have different advantages and disadvantages, it is becoming more and more difficult to select an appropriate convolutional neural network model in an actual fish identification project. The identification of underwater fish is a challenge task due to varies in illumination, low contrast, high noise, low resolution and sample imbalance between each type of image from the real underwater environment. In addition, deploying models to mobile devices directly will reduce the accuracy of the model sharply. In order to solve the above problems, Fish Recognition Ground-Truth dataset was used to training model in this study, which is provided by Fish4Knowledge project from University of Edinburgh. It contains 27,370 images with 23 fish species, and has been labeled manually by marine biologists. AlexNet, GoogLeNet, ResNet and DenseNet models were selected initially according to the characteristics of real-time underwater fish identification task, then a comparative experiment was designed to explore the best network model. Random image flipping, rotation and color dithering were used to enhance data based on ground-truth fish dataset in response to the limited number of underwater fish images. Considering that there was a serious imbalance in the number of samples in each category, the label smoothing technology was used to alleviate model overfitting. The Ranger optimizer and Cosine learning rate attenuation strategy were used to further improve the training effect of the models. The accuracy and recall rate information of each model were recorded and counted. The results showed that, the accuracy and recall rate of the fish recognition model based on DenseNet reached 99.21% and 96.77% in train set and validation set respectively, its F1 value reached 0.9742, which was the best model obtained in the experiment. Finally, a remote fish identification system was designed based on Python language, in this system the model was deployed to linux server and the Android APP was responsible for uploading fish images via http to request server to identify the fishes and displaying the identification information returned by server, such as fish species, profiles, habits, distribution, etc. A set of recognition tests were performed on real Android phone and the results showed that in the same local area net the APP could show fish information rapidly and exactly within 1 s.</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">fish identification model</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">cnn</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">model evaluation</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">android</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">ground-truth</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">real-time identification system</subfield></datafield><datafield tag="653" ind1=" " ind2="0"><subfield code="a">Agriculture (General)</subfield></datafield><datafield tag="653" ind1=" " ind2="0"><subfield code="a">Technology (General)</subfield></datafield><datafield tag="700" ind1="0" ind2=" "><subfield code="a">YANG Ling</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="0" ind2=" "><subfield code="a">YU Huihui</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="0" ind2=" "><subfield code="a">CHEN Yingyi</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="773" ind1="0" ind2="8"><subfield code="i">In</subfield><subfield code="t">智慧农业</subfield><subfield code="d">Editorial Office of Smart Agriculture, 2021</subfield><subfield code="g">4(2022), 1, Seite 130-139</subfield><subfield code="w">(DE-627)DOAJ078619246</subfield><subfield code="x">20968094</subfield><subfield code="7">nnns</subfield></datafield><datafield tag="773" ind1="1" ind2="8"><subfield code="g">volume:4</subfield><subfield code="g">year:2022</subfield><subfield code="g">number:1</subfield><subfield code="g">pages:130-139</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doi.org/10.12133/j.smartag.SA202202006</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doaj.org/article/6ee1de60dd7f467384128fea694724ad</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">http://www.smartag.net.cn/CN/10.12133/j.smartag.SA202202006</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="2"><subfield code="u">https://doaj.org/toc/2096-8094</subfield><subfield code="y">Journal toc</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_USEFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SYSFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_DOAJ</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SSG-OLC-PHA</subfield></datafield><datafield tag="951" ind1=" " ind2=" "><subfield code="a">AR</subfield></datafield><datafield tag="952" ind1=" " ind2=" "><subfield code="d">4</subfield><subfield code="j">2022</subfield><subfield code="e">1</subfield><subfield code="h">130-139</subfield></datafield></record></collection>
|
| score |
7.4014015 |