Image fine-grained classification method, system and equipment

Abstract

The invention discloses an image fine-grained classification method, system and equipment, and the method comprises the steps: carrying out the preprocessing and normalization processing of an image, inputting the image into a convolutional neural network, and carrying out the feature extraction, thus the conventional image spatial features are considered and the features of an image channel are considered, the fine-grained image classification effect can be improved; then obtaining a proportional relation between different channels through an SENet module, and identifying more fine discriminative regions of the image; adopting a clustering method of metric learning to find a plurality of attention areas in one image, so that the distance of the local features of the same category in the image is further shortened, and the distance of the local features of different categories in the image is further increased; and finally, training the image classification model continuously until the loss function converges to obtain an optimal image classification model, so that the images are classified through the model, and the technical problem of poor accuracy of fine-grained classification of the images in the prior art is solved.

Description

technical field [0001] The present application relates to the technical field of image processing, and in particular to a method, system and device for fine-grained classification of images. Background technique [0002] The purpose of image fine-grained classification is to distinguish the subordinate classes of a common class, such as distinguishing swans, wild geese, mandarin ducks, pelicans, seagulls, etc. from birds. The subclasses of these birds basically have similar global Geometric features and appearance features, this fine-grained classification is different from the ordinary classification between cats and dogs. Fine-grained classification needs to be able to find the nuances of the object, and extract these nuances into the features of the object, so as to realize the further classification of the object. In particular, finding the most discriminative regional features as much as possible in different parts of the image is crucial to the effect of classificatio...

AI Technical Summary

Problems solved by technology

[0005] This application provides a method, system and device for fine-grained classification of images, which is used to solve the problem of poor accuracy of fine-grained classification of images in the prior art technical problem

Images