Segmenting Freshwater Fish Images with Convolutional Neural Networks

• Nicolas Figueiredo Cavalcante Sales UNIR
• Carolina Yukari Veludo Watanabe UNIR https://orcid.org/0000-0002-6237-2323

DOI: https://doi.org/10.5753/wsis.2024.33668

Resumo

The acquisition of important information from fisheries for studies in Ichthyology and Fisheries currently takes place manually. Data on the length and morphology of freshwater fish are valuable for researchers to determine indicators in their studies. Technological development in this area aims to increase the speed and effectiveness of information gathering, assisting researchers, students, and fishermen in this goal. Aligned with this objective, this work presents a comparison between two models, Mask R-CNN and YOLOv8, used to segment fish images and generate their masks, ultimately feeding them into an automatic measurement algorithm, which is the broader purpose of this project. The results show that the models can segment various types of fish in different positions and environments, with Mask R-CNN achieving 80.16% using the IOU (Intersection over Union) metric and YOLOv8 achieving 86.15%.

Palavras-chave: Image segmentation, Deep Learning, Convolutional Neural Networks, Freshwater fish image segmentation, Mask R-CNN, YOLO, YOLOv8

Referências

CHOLLET, François. Deep learning with Python. Second edition. Shelter Island: Manning Publications, 2021.

GOOGLE. Google Colaboratory. 2024. Disponível em: [link].

HE, Kaiming et al. Mask R-CNN. Versão arXiv:1703.06870. [S. l.]: arXiv, 24 jan. 2018. arXiv:1703.06870 [cs]. Disponível em: [link]. Acesso em: 14 ago. 2024.

JOCHER, Glenn; CHAURASIA, Ayush; QIU, Jing. Ultralytics YOLOv8. Versão 8.0.0. 2023. Disponível em: [link].

LIN, Tsung-Yi et al. Microsoft COCO: Common Objects in Context. Versão arXiv:1405.0312. [S. l.]: arXiv, 20 fev. 2015. arXiv:1405.0312 [cs]. Disponível em: [link]. Acesso em: 14 ago. 2024.

REDMON, Joseph et al. You Only Look Once: Unified, Real-Time Object Detection. Versão arXiv:1506.02640. [S. l.]: arXiv, 9 maio 2016. arXiv:1506.02640 [cs]. Disponível em: [link]. Acesso em: 14 ago. 2024.

ROCHA, Wan Song et al. Automatic measurement of fish from images using convolutional neural networks. Multimedia Tools and Applications, [s. l.], 18 abr. 2024. Disponível em: [link]. Acesso em: 14 ago. 2024. DOI: 10.1007/s11042-024-19180-1

ROCHA, Wan Song; DORIA, Carolina Rodrigues Da Costa; WATANABE, Carolina Yukari Veludo. Fish Detection and Measurement based on Mask R-CNN. In: CONFERENCE ON GRAPHICS, PATTERNS AND IMAGES, 2020, Brasil. Anais Estendidos da Conference on Graphics, Patterns and Images (SIBRAPI Estendido 2020). Brasil: Sociedade Brasileira de Computação, 7 nov. 2020. p. 183–186. Disponível em: [link]. Acesso em: 14 ago. 2024.

SOHAN, Mupparaju; SAI RAM, Thotakura; RAMI REDDY, Ch. Venkata. A Review on YOLOv8 and Its Advancements. In: JACOB, I. Jeena; PIRAMUTHU, Selwyn; FALKOWSKI-GILSKI, Przemyslaw (org.). Data Intelligence and Cognitive Informatics. Singapore: Springer Nature Singapore, 2024. (Algorithms for Intelligent Systems). p. 529–545. E-book. Disponível em: [link]. Acesso em: 14 ago. 2024.

WADA, Kentaro. Labelme: Image Polygonal Annotation with Python. [s. d.]. Disponível em: [link].

WU, Yuxin et al. Detectron2. 2019. Disponível em: [link].