Comprehensive smartphone image dataset for fish species identification in Bangladesh's freshwater ecosystems
Saieef Sunny, Shiam Prodhan, Nazmuj Shakib, Mohammad Rifat Ahmmad Rashid, Nafees Mansoor
Journal Article. Data in Brief, May 12, 2025, p. 111629 (2025).
Abstract
The Bangladesh Fish Species Identification dataset encompasses 24,925 images representing 21 freshwater species commonly found in the rivers and ponds of Bangladesh. Photographs were taken using smartphones under carefully controlled conditions to ensure clarity and proper framing, and they include variations in lighting and background for realistic depiction. Species covered in this dataset include Aair, Boal, Chapila, DeshiPuti, Foli, Ilish, KalBaush, Katla, Koi, Magur, Mrigel, Pabda, Pangas, Puti, Rui, Shol, Shorputi, Taki, Tarabaim, Telapiya, and Tengra. By providing an extensive collection of labeled, high-quality images, this dataset serves as a valuable foundation for research on aquatic biodiversity, fishery management, and the development of precise machine-learning models for species recognition. It supports investigations into morphological variations, facilitates the evaluation of deep learning techniques, and assists in understanding fish distribution across diverse habitats. © 2025 The Author(s)
Keywords
Bangladesh, Biodiversity, Dataset, Deep learning, Fish species identification, Freshwater fish, Image recognition, Machine learning