Fine-grained vehicle image classification algorithm based on improved Smaller VGGNet
To improve the ability of vehicle category information recognition and analysis in the field of intelligent transportation,a fine-grained vehicle image classification algorithm model based on improved SmallerVGGNet is proposed.The model first readjusted the neural network structure to improve training stability and classification accuracy.Secondly,replace the filling method of features to better capture local information and alleviate overfitting.Finally,ELU is used to replace the original ReLU activation function to accelerate the convergence speed of the model.On a publicly available dataset,the improved model with six existing image classification algorithms were compared,experimental results show that the improved algorithm outperforms significantly the compared algorithms and exhibits better stability during the training process.
万方数据
