__Citation__ This dataset has been introduced in: Jenni Raitoharju, Ekaterina Riabchenko, Iftikhar Ahmad, Alexandros Iosifidis, Moncef Gabbouj, Serkan Kiranyaz, Ville Tirronen, Johanna Ärje, Salme Kärkkäinen, Kristian Meissner, "Benchmark database for fine-grained image classification of benthic macroinvertebrates", Image and Vision Computing, Vol. 78, 2018. [https://doi.org/10.1016/j.imavis.2018.06.005](https://doi.org/10.1016/j.imavis.2018.06.005) __Introduction__ This is a dataset for automatic fine-grained classification of benthic macroinvertebrates. There are 15,074 images from 64 categories. The number of images per category varies from 577 to 7. __Contents__ For each image, we provide: -Original image, varying size, png -Cropped image scaled to 256x256 pixels, png -Object mask for the cropped image 256x256, png -30 different feature vectors (one for each of partition of each dataset) of 4096 elements generated by CNN networks The database contains 3 different datasets: _Dataset 1:_ -All 15,074 images -64 categories -Images per category: 7-577 _Dataset 2:_ -11,832 images -29 categories (1-29) -Images per category: 230-577 _Dataset 3:_ -3,272 images -9 categories (19-27) -Images per category: 322-395 Every dataset has been divided into 10 different partitions: -50 % for training -20% for validation -30% for testing -The partitions are given in .txt file separately for each dataset

2021