SimCLR-Enhanced MammalHairNet: Advancing Species Identification Through Hair Scale Classification

ori_all_training.tar Contains all raw data outside the test set 2025_85_720[_per**}.tar The dataset comprises data excluded from the test set, which have been processed through SAM-based semantic segmentation followed by a series of augmentation techniques, including random contrast adjustment, cropping, noise addition, mirroring, and rotation. 2025_15_100.tar Test set seg.tar Semantic Segmentation Models and Annotated Datasets model_resnet_efficientnet_vit.zip Models trained by transfer learning model_resnet_efficientnet_vit_from_scratch.zip Models trained from scratch model_fintuning.zip Models trained through SimCLR and fine-tuning