Deep Sea Vision: Enhanced YOLO-Driven Framework for High-Precision Marine Fish Detection

Fuente: Zenodo
Saved in:
Bibliographic Details
Main Authors: Kanneboina Ashok, Mohaideen A, Manivannan.S, Radhakrishnan M, M. Shailaja
Format: Recurso digital
Published: Zenodo 2026
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866901801140224000
author Kanneboina Ashok
Mohaideen A
Manivannan.S
Radhakrishnan M
M. Shailaja
author_facet Kanneboina Ashok
Mohaideen A
Manivannan.S
Radhakrishnan M
M. Shailaja
contents <p><span class="fontstyle0">This work aims at studying the species detection of marine animals with deep learning approaches to<br>YOLOv9 and Faster R-CNN algorithms. The main goal is to identify seven categories of species of underwater objects:<br>Fish, Jellyfish, Penguins, Puffins, Sharks ,Starfish, and Stingrays. Here, the study combines the high detection precision<br>of YOLOv9 with 92% with the more robust performance of the Faster R-CNN with 89% to fashion an all-inclusive<br>detection system fit for any underwater environment. For training and testing purposes, a well-selected dataset was used<br>that included a good variety of species to consider every possibility. It can be noted that the use of the proposed methods<br>affords a significant improvement of recognition accuracy and increasing the rate of detection. In many ways, this<br>research suggests that the potential for use of these technologies in such fields as marine biology, and specifically,<br>documenting and conserving marine environments. Thus, using these optimized detection methods, researchers and<br>conservationists can obtain the necessary data on marine diversity and develop the protection of underwater habitats.<br>Finally, this work acknowledges the importance of computer vision in augmenting knowledge of marine life forms.</span> </p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_19588551
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle Deep Sea Vision: Enhanced YOLO-Driven Framework for High-Precision Marine Fish Detection
Kanneboina Ashok
Mohaideen A
Manivannan.S
Radhakrishnan M
M. Shailaja
<p><span class="fontstyle0">This work aims at studying the species detection of marine animals with deep learning approaches to<br>YOLOv9 and Faster R-CNN algorithms. The main goal is to identify seven categories of species of underwater objects:<br>Fish, Jellyfish, Penguins, Puffins, Sharks ,Starfish, and Stingrays. Here, the study combines the high detection precision<br>of YOLOv9 with 92% with the more robust performance of the Faster R-CNN with 89% to fashion an all-inclusive<br>detection system fit for any underwater environment. For training and testing purposes, a well-selected dataset was used<br>that included a good variety of species to consider every possibility. It can be noted that the use of the proposed methods<br>affords a significant improvement of recognition accuracy and increasing the rate of detection. In many ways, this<br>research suggests that the potential for use of these technologies in such fields as marine biology, and specifically,<br>documenting and conserving marine environments. Thus, using these optimized detection methods, researchers and<br>conservationists can obtain the necessary data on marine diversity and develop the protection of underwater habitats.<br>Finally, this work acknowledges the importance of computer vision in augmenting knowledge of marine life forms.</span> </p>
title Deep Sea Vision: Enhanced YOLO-Driven Framework for High-Precision Marine Fish Detection
url https://doi.org/10.5281/zenodo.19588551