Automatic labeling of fish species using deep learning across different classification strategies

  1. Jareño, Javier 2
  2. Bárcena-González, Guillermo 2
  3. Castro-Gutiérrez, Jairo 1
  4. Cabrera-Castro, Remedios 1
  5. Galindo, Pedro L. 2
  1. 1 Biology Department, University of Cádiz, Cádiz, Spain
  2. 2 Computer Science Department, University of Cádiz, Cádiz, Spain
Revista:
Frontiers in Computer Science

ISSN: 2624-9898

Año de publicación: 2024

Volumen: 6

Páginas: 1-12

Tipo: Artículo

DOI: 10.3389/FCOMP.2024.1326452 GOOGLE SCHOLAR lock_openAcceso abierto editor

Otras publicaciones en: Frontiers in Computer Science

Objetivos de desarrollo sostenible

Resumen

Convolutional neural networks (CNNs) have revolutionized image recognition.Their ability to identify complex patterns, combined with learning transfertechniques, has proven eective in multiple fields, such as image classification. Inthis article we propose to apply a two-step methodology for image classificationtasks. First, apply transfer learning with the desired dataset, and subsequently, ina second stage, replace the classification layers by other alternative classificationmodels. The whole methodology has been tested on a dataset collected atConil de la Frontera fish market, in Southwest Spain, including 19 dierentfish species to be classified for fish auction market. The study was conductedin five steps: (i) collecting and preprocessing images included in the dataset,(ii) using transfer learning from 4 well-known CNNs (ResNet152V2, VGG16,EcientNetV2L and Xception) for image classification to get initial models, (iii)apply fine-tuning to obtain final CNN models, (iv) substitute classification layerwith 21 dierent classifiers obtaining multiple F1-scores for dierent trainingtest splits of the dataset for each model, and (v) apply post-hoc statisticalanalysis to compare their performances in terms of accuracy. Results indicatethat combining the feature extraction capabilities of CNNs with other supervisedclassification algorithms, such as Support Vector Machines or Linear DiscriminantAnalysis is a simple and eective way to increase model perfomance.

Referencias bibliográficas

  • Allken, (2019), ICES J. Mar. Sci, 76, pp. 342, 10.1093/icesjms/fsy147
  • Barbedo, (2019), Biosyst. Eng, 180, pp. 96, 10.1016/j.biosystemseng.2019.02.002
  • Bonferroni, (1936), Pubbl. R. Ist. Super. di Sci. Econom. Commer. Firenze, 8, pp. 3
  • Buslaev, (2020), Information, 11, pp. 125, 10.3390/info11020125
  • Chollet, (2017), Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 10.1109/CVPR.2017.195
  • Cutler, (2007), Ecology, 88, pp. 2783, 10.1890/07-0539.1
  • Deep, (2019), 2019 6th International Conference on Signal Processing and Integrated Networks (SPIN), pp. 665, 10.1109/SPIN.2019.8711657
  • Dobeson, (2016), Econ. Soc, 45, pp. 454, 10.1080/03085147.2016.1224143
  • Duncan, (1955), Biometrics, 11, pp. 1, 10.2307/3001478
  • Dunnett, (1955), J. Am. Stat. Assoc, 50, pp. 1096, 10.1080/01621459.1955.10501294
  • Franco, (1990), Analyt. Chim. Acta, 233, pp. 143, 10.1016/S0003-2670(00)83471-6
  • Galindo, (2000), Advances in Pattern Recognition: Joint IAPR International Workshops SSPR 2000 and SPR 2000 Alicante, Spain, August 30-September 1, 2000 Proceedings, pp. 796, 10.1007/3-540-44522-6_82
  • Games, (1979), Psychol. Bull, 86, pp. 978, 10.1037//0033-2909.86.5.978
  • Goodfellow, (2016), Deep Learning
  • He, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 10.1109/CVPR.2016.90
  • He, Computer Vision-ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11–14, 2016, Proceedings, Part IV, pp. 630
  • Huang, (2017), Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 10.1109/CVPR.2017.243
  • Hussain, (2019), Advances in Computational Intelligence Systems: Contributions Presented at the 18th UK Workshop on Computational Intelligence, September 5–7, 2018, Nottingham, UK, pp. 191, 10.1007/978-3-319-97982-3_16
  • Ibraheam, (2021), AI, 2, pp. 552, 10.3390/ai2040034
  • Jarek, (2019), Central Eur. Bus. Rev, 8, pp. 46, 10.18267/j.cebr.213
  • Kaya, (2019), Comput. Electr. Agric, 158, pp. 20, 10.1016/j.compag.2019.01.041
  • Knauer, (2019), Rem. Sens, 10.3390/rs11232788
  • Krizhevsky, (2012), Advances in Neural Information Processing Systems
  • Luan, (2020), Fisher. Res, 227, pp. 105534, 10.1016/j.fishres.2020.105534
  • Montalbo, (2019), 2019 IEEE 9th International Conference on System Engineering and Technology (ICSET), pp. 396, 10.1109/ICSEngT.2019.8906433
  • Munoz, (2013), Stochastic Environ. Res. Risk Assess, 27, pp. 1171, 10.1007/s00477-012-0652-3
  • Norouzzadeh, (2018), Proc. Natl. Acad. Sci, 115, pp. E5716, 10.1073/pnas.1719367115
  • Nuraini, (2022), IJICS, 6, pp. 147, 10.30865/ijics.v6i3.5565
  • Palmer, (2022), Fisher. Res, 246, pp. 106166, 10.1016/j.fishres.2021.106166
  • Pizarro, (2002), Neurocomputing, 48, pp. 155, 10.1016/S0925-2312(01)00653-1
  • Pundlik, (2016), 2016 7th International Conference on Intelligent Systems, Modelling and Simulation (ISMS), pp. 120, 10.1109/ISMS.2016.57
  • Saberioon, (2018), Sensors, 18, pp. 1027, 10.3390/s18041027
  • Seabold, (2010), Proceedings of the 9th Python in Science Conference, pp. 10, 10.25080/Majora-92bf1922-011
  • Shang, (2018), 2018 10th International Conference on Wireless Communications and Signal Processing (WCSP), pp. 1, 10.1109/WCSP.2018.8555591
  • Comparison of post hoc tests for unequal variance2233 ShingalaM. C. RajyaguruA. Int. J. New Technol. Sci. Eng22015
  • Shorten, (2019), J. Big Data, 6, pp. 1, 10.1186/s40537-019-0197-0
  • Simonyan, (2014), arXiv preprint arXiv:1409.1556
  • Szegedy, (2014), Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 10.1109/CVPR.2015.7298594
  • Tan, (2021), International Conference on Machine Learning, pp. 10096
  • Tukey, (1949), Biometrics, 5, pp. 99, 10.2307/3001913