IAS-16k-C

IAS-16k-C (Invasive Alien Species 16k-Class) is a large-scale benchmark dataset for fine-grained visual classification at the intersection of computer vision and ecological conservation, dedicated to the identification of highly harmful invasive alien plant species worldwide. It contains a total of 162,698 high-quality images covering 51 representative invasive alien species, specifically constructed for the training, validation, and testing of deep learning models, and supports high-precision fine-grained species classification tasks. The dataset is strictly partitioned into three subsets at a ratio of 7:2:2 to ensure fair and reliable model evaluation: Training set: 113,860 images (70.0%) Validation set: 24,381 images (15.0%) Test set: 24,457 images (15.0%)