Lightweight visual information extraction method using grouping wavelet packet transformation

The invention discloses a lightweight visual information extraction method using packet wavelet packet transformation, and relates to general image data processing or generation. A convolution operation in the convolutional neural network model is replaced by a packet wavelet packet transformation module, and the input feature map is transformed from a spatial domain to a frequency domain to obtain a frequency domain feature map; performing depth separable convolution on the frequency domain feature map by using a convolution kernel in a frequency domain, and adding and splicing output results to obtain a target frequency domain feature map; restoring the target frequency domain feature map to a spatial domain by using a grouping wavelet packet inverse transformation module to obtain an output feature map; and carrying out forward propagation of a convolutional neural network model on the output feature map, and carrying out iterative training by using a gradient descent method to obtain an image classificati