Fine-grained image classification method and system

The invention discloses a fine-grained image classification method and a fine-grained image classification system. The fine-grained image classification method specifically comprises the following steps: initializing a hierarchical visual converter network by using pre-training parameters; sampling the data; performing underlying feature comparative learning according to the sampled data to obtain first prediction loss; performing deep feature balance processing according to the sampled data to obtain second prediction loss; obtaining the overall loss according to the first prediction loss and the second prediction loss; network parameters are updated through back propagation; repeating the above steps until the parameters of the hierarchical visual converter network converge, and storing the updated network parameters; initializing the hierarchical visual converter network by using the stored network parameters; and carrying out image category prediction according to the initialized hierarchical visual conve