Convolutional neural network model and data processing method and device

Abstract

The invention relates to a convolutional neural network model, a data processing method and device, relating to the technical field of deep learning. The convolutional neural network model comprises an input layer, at least one first convolutional layer, a feature fusion layer and an output layer, the first convolutional layer comprises a deep convolutional sub-layer and a point-by-point convolutional sub-layer, the deep convolutional sub-layer comprises m channel groups, and each channel group comprises at least two input channels and at least two output channels. Through the first convolution layer, the data features input to the plurality of input channels are divided into a plurality of groups, and the data features input to each input channel in each group can be shared by the outputchannels in the same group, so that the data feature extraction in the convolution process is more comprehensive, and the accuracy of the convolution neural network model is improved.

Description

technical field [0001] The invention relates to the technical field of deep learning, in particular to a convolutional neural network model, a data processing method and a device. Background technique [0002] In recent years, Convolutional Neural Network (CNN) has achieved rapid development and has achieved remarkable results in areas such as image recognition. [0003] In related technologies, in order to reduce the complexity of the convolutional neural network and improve computational efficiency, the convolutional layer in the convolutional neural network can be divided into a deep convolution sublayer and a convolutional kernel with a convolution kernel size of h×w A pointwise convolutional sublayer of size 1×1, where, figure 1 A schematic diagram of the structure of the convolutional layer is shown. [0004] Such as figure 1 As shown, the depth convolution sublayer contains several sets of one-to-one corresponding input channels and output channels. The output chan...

AI Technical Summary

Problems solved by technology

[0006] The embodiment of the present invention provides a convolutional neural network model, data processing method and device, which can solve the problem that one output channel of the deep convolution sublayer in the related art is only affected by the data characteristics in one input channel, causing the convolution process to The extraction of data features in is not comprehensive enough, which affects the accuracy of the convolutional neural network. The technical solution is as follows:

Images