Research On Image Classification Algorithms Based On Convolutional Neural Network

Image classification refers to an image processing method that distinguishes different types of targets based on different characteristics reflected in image information.With the further development of related research of intelligent optimization theory,deep learning has become a research hotspot for image classification.Many researchers have conducted extensive research on the theory and methods of deep learning and achieved considerable progress,but the slow convergence rate and the classification accuracy in deep learning theory still need to be further improved.It is one of the important research directions to construct better performance deep learning algorithms and use in image classification.This thesis mainly studies the image classification algorithm based on convolutional neural network.The main contents are as follow: 1.Aiming at the problem that the convolutional neural network has slow convergence rate in image classification,an improved image classification algorithm for convolutional neural network is proposed.The genetic algorithm is used to improve the weight update mode of convolutional neural network and effectively improve convergence speed and classification accuracy.2.Convolutional neural networks usually improve the classification accuracy by increasing the depth of the network,but the increase in the number of network layers will make it difficult to train the network and even reduce the classification accuracy.In order to solve this problem,an image classification method based on ensemble learning Bagging and convolutional neural network is proposed.This method,improve the accuracy of image classification,uses the convolutional neural network as the based classifier and polls the convolutional neural network without increasing the number of network layers.3.For the voting mechanism in ensemble learning Bagging algorithm does not consider the difference between the performance of the base classifier,resulting in poor classification results,an image classification method based on the improved Bagging voting weight strategy is proposed.It takes the out-of-data as the test set's accuracy rate as the voting weight of the convolutional neural network based classifier,the voting weight of the strong classifier is improved.The experimental results verify the effectiveness of the method.