BDryFish: An Image Dataset for Dried Fish Classification in Bangladesh

BDryFish is a curated image dataset comprising 5,874 high-quality images of dried fish from 22 distinct species commonly found and consumed in Bangladesh. Each species is represented in two packaging formats — Single (individual fish) and Bulk (grouped fish) — to capture real-world variability in fish presentation. The dataset is designed to support AI and computer vision research, particularly in areas such as:Fish species classificationAgricultural and fisheries product recognitionFood packaging and supply chain monitoringObject detection and segmentationDataset Summary:Total images: 5,874Single images: 3,026Bulk images: 2,848Fish categories: 22 species × 2 formats (Single, Bulk)Formats: Standard image files (e.g., JPEG/PNG)Fish Species Examples:Banded GouramiBombay DuckHilsaGiant Largehead HairtailIndo-Pacific King MackerelSun-Dried Swamp BarbWhite Sardinella...and more.Use Cases:This dataset is ideal for building image classification models, training deep learning pipelines, evaluating transfer learning architectures (e.g., MobileNet, EfficientNet), and conducting domain-specific research in fishery management and food safety.