Target classification method based on adaptive feedforward neural network

Abstract

The invention discloses a target classification method based on an adaptive feedforward neural network. The method is characterized by comprising the following steps: constructing a feedforward neural network model, namely reconstructing a nonlinear feedforward neural network system into a linear weighted nonlinear system through a Taylor series expansion; performing weight updating by taking a weight estimation error as a drive, namely extracting a weight estimation error information variable according to a constructed auxiliary variable, and correspondingly designing an adaptive rate of weight updating; and carrying out target classification, namely classifying a target image according to the weight-updated feedforward neural network model. According to the invention, through the improved weight adaptive estimation method, the limitation of a traditional neural network weight updating algorithm is solved, and the rapid recognition of a target in an image is realized.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to an object classification method based on an adaptive feedforward neural network. Background technique [0002] Target classification can realize target recognition in images and provide information data support for decision-making systems. In recent years, deep learning methods have been extensively studied for object classification, but traditional object classification algorithms are not ideal in terms of accuracy and robustness. Most of the existing results use gradient algorithms to minimize output errors to derive classic neural network weights. The value update law, its slow convergence process will lead to a large amount of data and time training of the neural network, increasing the time cost and consumption cost of neural network learning. A new approach is needed to solve the problem of neural networks. Contents of the invention [0003] The purpose of the pre...

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide a target classification method based on an adaptive feed-forward neural network to solve the traditional neural network convergence problem limited by the traditional gradient algorithm, as well as a large amount of data training and Technical issues that consume time and cost

Images