Depth scattering convolution network learning method and system based on nuclear space

Abstract

The invention relates to a depth scattering convolution network learning method and system based on a nuclear space. The system comprises a parameterization wavelet generation module, a multinuclear learning module, a scattering decomposition module and a supporting vector machine classification module. The parameterization wavelet generation module constructs a wavelet filter group via a randomized parameter. The multinuclear learning module carries out convolution decomposition of a scattering network on a training data set based on the above filter group, separately projects decomposition coefficients of convolution paths to a nuclear feature space, and selects an optimal convolution path by use of a multinuclear learning algorithm. The scattering decomposition module carrying out scattering decomposition on a test data set based on the optimal convolution path. The obtained decomposition coefficients are classified by the support vector machine classification module. According to the invention, classification accuracy of all types of digital images can be effectively improved.

Description

technical field [0001] The invention relates to a digital image classification scheme, in particular to a kernel space-based deep scattering convolution network learning method and system. Background technique [0002] The classification of digital image signals is usually solved jointly by feature extraction operators and classifiers. The ability of features extracted by feature extraction operators to identify intra-class similarities and distinguish between-class differences is called feature separability, and this property obviously affects classification accuracy. When the feature extraction operator is determined, the feature separability is affected by the back-end classifier. In order to further improve the classification accuracy of the signal, a scheme is to jointly optimize the feature extraction operator and the classifier, which can provide the robustness of classifying different types of digital images. [0003] After searching the literature of the prior art...

AI Technical Summary

Problems solved by technology

However, this method only solves the learning problem of a single-layer filter bank, and the obtained signal decomposition coefficients do not contain interactive information between different scales and directions, which will eventually lead to inaccurate classification of image signals

Images