Self Supervised Image Classification

The self-supervised image classification task aims to obtain high-quality image representations through self-supervised learning methods and evaluate them by training a linear classifier on top. Self-supervised learning involves solving a pre-training task to learn representations, typically using specific loss functions such as contrastive loss to measure the similarity of sample pairs in the representation space. This task has significant application value in computer vision, effectively reducing the need for labeled data and enhancing the model's generalization ability.