基于TensorFlow卷积神经网络的图像识别

TensorFlow是Google公司发布的开源人工智能深度学习框架,卷积神经网络是进行图像识别的一种有效方法。本文在研究Tensorflow深度学习框架以及卷积神经网络的基础上,利用keras官方下载的cifar数据集,采用LeNet-5算法对数据进行了处理、建模、训练、并对模型进行了评估以及保存,利用测试集完成测试后,不同图像识别的准确率有所不同,青蛙识别的准确率最高,为79%,汽车的识别准确率为78%,猫和狗的识别准确率最低,分别为41%和53%,所有图像识别的平均准确率为65%。

Image Recognition Based on the Convolution Neural Network of Tensor Flow

TensorFlow which is an open source arti cial intelligence in-depth learning framework developed by Google.The convolution neural network is an effective method for image recognition.Based on the study of TensorFlow and convolution neural network,this paper uses the cifar data set downloaded from keras and uses lenet-5 algorithm to process,model,train,then the paper store the model and evaluate the model.After completing the test using the test set,the accuracy of different image recognition varies that frogs has the most accurate at 79%,cars is 78%,and cats and dogs are the least accurate at 41%and 53%,respectively.The average accuracy of all image recognition is 65%.