Image classification method based on improved ResNet network

Abstract

The invention provides an image classification method based on an improved ResNet network. The method comprises the steps of obtaining an image to be classified and inputting the image to be classified into the improved ResNet network. According to the invention, through carrying out multi-feature fusion on the input image, the image features can be further extracted, and the performance of the model is improved; meanwhile, through feature reuse of each residual block, the feature extraction effect is further improved; besides, by designing a new activation function, the problem of gradient explosion can be effectively solved under the condition that the characteristic value is relatively large. Compared with the existing ResNet network, the improved ResNet network provided by the invention can realize higher classification precision and higher classification speed.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an image classification method based on an improved ResNet network. Background technique [0002] Convolutional Neural Networks have long been one of the core algorithms in the field of image recognition and have a stable performance when learning data is sufficient. For general large-scale image classification problems, convolutional neural networks can be used to construct hierarchical classifiers, and can also be used to extract discriminative features of images in fine-grained recognition for other classifiers. study. Existing convolutional networks include LetNet, AlexNet, VGG, Google's Inception series and ResNet. Among them, the ResNet network can build an ultra-deep network structure through the residual (residual) structure to solve the problem of traditional convolutional neural networks. The problem of gradient disappearance or gradient explosion caused by the deepeni...

AI Technical Summary

Problems solved by technology

Existing convolutional networks include LetNet, AlexNet, VGG, Google's Inception series, and ResNet. Among them, the ResNet network can build an ultra-deep network structure through a residual (residual) structure, solving the problem of traditional convolutional neural networks. The problem of gradient disappearance or gradient explosion caused by deepening depth has a good effect on alleviating the degradation problem of deep network, but in the garbage classification scene, the existing ResNet network is not satisfactory in classifying pictures.

Images