FishAI: A Vision Transformer-based Tool for Image Classification of Marine Fish



Introduction

We collected raw data, performed data cleaning on the raw data and obtained a WoRMS (global data source) based dataset of fish images.
In this study, we present a deep learning based approach to classify fish from images, where vision transformer is used to extract high-level features, and Grad-CAM is applied to explain the patterns.


Availability: The FishAI is available at www.csbio.sjtu.edu.cn/bioinf/FishAI.




Figure 1. The flowchart of the proposed FishAI  

© 2017 Computational Systems Biology/Shen Group.