摘要：This article mainly introduces the image classification algorithm research based on the improved convolution neural network and some improvement ideas for the research of the classification based on the convolution neural network. The VGG16 model was improved in this article. By adding a dropout layer and a feature extraction layer, and performing L2 regularization on the loss function at the end, it deepens the model depth and improves the fit of the entire model. The experimental results show that the improved model can greatly improve the detection accuracy rate 98.7% than the traditional CNN algorithm 96.97% on the cat and dog data set.
关键词：Deep LearningImage ClassificationConvolutional Neural NetworkVGG16 Model
