Study on fish target detection method based on deep learning and its comparison with gray value method in an impurity experimental environment

Description

This dataset showcases the model training data used in the article, along with the raw data employed for variability analysis. However, it excludes fish behavior videos, as the behavioral videos utilized in this study were generated from other, as-yet unpublished research projects.

Steps to reproduce

The deep learning algorithm used in this study is YOLOv5, and the relevant target detection process was conducted in a Python 3.8 environment. Other configuration environment parameters include torch 1.8.1+cu102, torchaudio 0.8.1, and torchvision 0.9.1+cu102. T收到he target detection process using the gray-scale method was performed in Matlab 2017b.

Institutions

Categories

Licence