Enhancing Fish Taxonomy with AI-Assisted Photographic Identification Tools

Abstract

Accurate species identification is the cornerstone of every taxonomic investigation and an essential component 
of biological research protocols. Biologists are calling for more efficient methods to handle the identifying need. 
Even people without solid experience in computer science may now independently apply state-of-the-art 
algorithms to their problems and datasets due to machine learning software that is getting easier to use. However, 
fundamental knowledge and some familiarity with the technology being used are still necessary. Around the 
world, efforts are being made to use data and automated processes to manage better and increase the efficiency 
of the fish industry, decrease the need for human labour in fishing and aquaculture, prevent diseases in fish farms, 
prevent pollution of the environment, and increase the number of fish available for human consumption. The most 
advanced fish identification (ID) technologies, including those that are still in the early phases of development, 
are evaluated in this research. It demonstrates how, in a variety of real-world scenarios, they could be the best 
option. The identification tools that were reviewed included the following: morphometrics (like IPez), scale and 
otolith morphology, genetic methods (like SNPs and barcodes), interactive electronic keys (like IPOFIS), 
taxonomic reference collections, image recognition software, dichotomous key-based field guides, the use of 
taxonomists and folk local experts, and hydroacoustic. It is anticipated to assist environmental administrators, 
fishery managers, and other end users in choosing the species identification instruments that best meet their needs.