Image recognition method based on deep learning and related device

Abstract

The invention discloses an image recognition method based on deep learning and a related device. The method comprises the following steps: obtaining a plurality of to-be-processed feature maps corresponding to a to-be-recognized image through feature extraction; and performing multiple rounds of first convolution operation on each to-be-processed feature map based on a preset first expansion factor, so as to determine a first total feature map corresponding to each to-be-processed image according to a feature recognition result obtained by each round of first convolution operation. And determining an identification result of the to-be-identified image through a decoder. The first expansion factors corresponding to each round of first convolution operation are different, so that feature recognition results obtained by each round of first convolution operation are in different receptive fields. And the feature recognition result of each round of first convolution operation is used as an input item of the next round, so that the last round of first convolution operation can obtain a first total feature map representing fusion of feature results under different receptive fields. Through the process, the parameter quantity of the neural network model can be greatly reduced, and the model convergence speed is improved.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular to an image recognition method and related devices based on deep learning. Background technique [0002] Image recognition technology is mostly based on deep learning image classification algorithms, using pooling layer or multi-layer convolution to extract the features of the image to be recognized. In order to improve classification accuracy, downsampling operations are often used in related technologies to expand the range of receptive fields, thereby increasing the semantic information of feature images. However, expanding the range of the receptive field will lead to the loss of feature image detail information. [0003] In order to solve the above problems, the method of feature pyramid is often used at present, and the features of the image to be recognized under different receptive fields are extracted through multiple convolution kernels, and then the feat...

AI Technical Summary

Problems solved by technology

Although this processing method can repair the detailed information of the feature map to a certain extent, it greatly increases the parameters of the neural network model, resulting in a decrease in the convergence speed of the neural network.

Images